Generative AI x Public Relations Textbook: 100 Case Studies

Introduction

■ A 'New Weapon' Is Emerging in the Field of Public Relations

The field of public relations is always at the center of 'words,' 'information,' and 'emotions.'

Writing press releases.
Creating internal newsletters.
Posting on social media.
Refining management messages.
Delivering every word carefully to internal and external audiences during crisis management.
Events, recruitment, IR, brand strategy...

The volume, speed, and responsibility of the information handled by PR professionals
are increasing year by year.

A 'New Partner' Called AI Has Arrived

In the midst of this, what suddenly appeared on the scene isgenerative AI.

'Can this actually be used?'
'Are we in an era where AI even writes our text?'
'But is this really language that resonates with people?'

PR professionals are facing AI in their daily work.

■ What I Want to Convey in 'Generative AI x Public Relations Textbook: 100 Case Studies'

Because there are few real-world examples available, it is difficult to visualize the practical application of generative AI in PR.
I want to provide hints to help you visualize and incorporate it, even if just a little.

Therefore, I have compiled this as 100 realistic case studies of 'PR x Generative AI.'

This is neither a theoretical book nor a technical manual.

'How should I use AI for this task tomorrow?'
I have created 'hints for the answer' to that question.

■ The Background Behind the Creation of 'Generative AI x Public Relations Textbook: 100 Case Studies'

As an author, I have been working on the front lines of PR support, media relations, and content design for over 20 years.

When AI first appeared, to be honest, this is what I thought:

However, as I continued to use it, I realized something:

Generative AI can produce text for you.
However, "which words to adopt," "what order to arrange them in," and "what emotions to convey" are the very essence of a PR professional's work.

In fact, it is precisely because AI exists that
what we in PR "truly need to do" becomes clearly visible.

"Generative AI x Public Relations Textbook: 100 Case Studies" was born from that realization.

■ How to use "Generative AI x Public Relations Textbook: 100 Case Studies"

• All 100 stories begin with "common problems and challenges in actual PR work."

• Each story is packed with "cases," "solutions," "AI tools and prompts used," "columns," "Q&A," "reader exercises," and "final thoughts."

• You can start reading from anywhere.

• How you read it is up to you.
  "Pick and choose stories based on the themes you are struggling with"
  "Use it as a collection of prompts"
  "Look at examples of how to use different AI tools"
  There is bound to be one story that hits home for some part of your work.

How to use "Generative AI x Public Relations Textbook: 100 Case Studies"

─ Please keep it on hand as "content to use," not "content to read."

■ Things you should know first

I wrote this to help you with your practical PR work.

In each of the 100 case studies, I have packed inconcrete examples of AI usage, prompts, and solutions.

Rather than "reading carefully from the beginning," it is fine to just
"pick and read only the stories close to the themes you are currently struggling with."
That level of commitment is perfectly fine.

■ Basic structure of this book

Every story is created following the flow below.

■ Introduction

The PR challenge addressed in that story is depicted asan episode of a PR person who could actually exist.If you read it and think, "Oh, I have this happen to me too..." then it is a success.

■ Case

How AI was actually used and in what order the work was approached.
The AI tools featured and prompt examples are all described in detail.

■ Resolution

What results were achieved,
what insights were gained,
and what were the key points?

To ensure it doesn't end with just 'doing it,' the learning points are clearly summarized.

■ Summary of AI Tools and Prompts Used

ChatGPT, Claude, NotebookLM, Genspark, Canva, Perplexity, Copilot, Notion, Gemini, etc.
For each tool used in the story, the actual prompts are included as-is for your reference.

It is designed so you can copy and paste them to try in your own workplace.

■ Column

Each story includes a short column tailored to the theme.
Topics like 'the essence of PR,' 'how to approach language,' and 'the sense of distance from AI'
address the common 'mental hurdles' practitioners often face.

■ What should I do in this situation?

A mini FAQ in Q&A format covering common real-world scenarios.
For each story, we provide answers to 'common points of confusion.'

■ Reader Work

Each story includes work assignments you can take back and use in your own tasks immediately.
They are designed with a focus on 'the first step you can take today.'

■ Conclusion

A summary of the story and how it connects to the next steps.

■ We also recommend these ways to read!

• Pick and choose themes that match your current concerns

• Read only the usage examples for specific AI tools in sequence

• Read the whole book and review everything with the prompt collection at the end

• Use it for in-house study sessions, practicing one story at a time as a team

• Use it as your own personal prompt collection, referencing it as needed

■ List of AI tools featured in this book (9 in total)

AI Tool Main Features
ChatGPT: Conversational generative AI. Ideal for text structure and drafting.
Gemini: Effective for multimodal analysis, planning support, and PR strategy.
Claude: Excels at natural tone refinement, long-form summarization, and simplifying language.
NotebookLM: Strong for organizing and analyzing internal documents and historical data.
Genspark: Useful for brainstorming, templating, and prompt generation.
Canva: Best for creating visual materials, banners, and posters.
Perplexity: Convenient for real-time information search, competitor research, and industry trend analysis. Copilot: Supports Excel/Word tasks and assists in creating tables and charts.
Notion: For information organization, project management, and AI-powered summarization and heading creation.

■ The Goal of "Generative AI x PR Textbook: 100 Case Studies"

Even as AI evolves,
it is you, the PR professional, who decides "who to tell, what to tell, and how to tell it."

"Generative AI x PR Textbook: 100 Case Studies" is a practical guide to becoming someone who can make informed decisions.

Please feel free to "use" it starting today,
from wherever you like, in whatever order you prefer.

"Generative AI x PR Textbook: 100 Case Studies" is a living, updated document.
Because generative AI evolves so rapidly, there is a high probability that the content will quickly become outdated.
I alreadyManuswant to add to the content.
Please be sure to check back regularly.

※ Please see above for the table of contents with clickable links.
Part 1 | Fundamentals and Design — Building the Foundation of PR
Chapter 1 | Structure and Expression — Refining "Effective Design"
01 | I can't write a press release! How should I convey this?
02 | Zero replies after sending to media... Reviewing selection and messaging
03 | Rebuilding "unread internal PR" through structure
04 | How to differentiate between IR materials and PR materials?
05 | How to organize and communicate multiple services?
06 | No reaction to SNS posts... What is the identity of a lost tone and manner?
07 | Has your owned media stopped updating?
08 | Designing for "empathy" rather than aiming for SNS "buzz"
09 | How much should PR write for FAQs and CS copy?
10 | Techniques for rediscovering "material" when press release ideas run dry

Chapter 2 | Verbalizing Strategy — Creating the "Core" of PR
11 | PR policy is not set! Initial strategy design
12 | "Are PR KPIs necessary?" Breaking away from vague evaluations
13 | The intersection of PR and IR: Information design that speaks to the future
14 | The big picture of IPO preparation PR and "timing of words"
15 | How to communicate governance? Techniques for expressing transparency
16 | How to verbalize a company's "reason for existence"
17 | The first step in startup PR
18 | How to translate management's "impossible demands" into PR?
19 | Internal dialogue techniques for when PR activities are not understood
20 | Can't get a PR budget... How to become someone who "can speak up"

Chapter 3 | Relationship Building and Empathy — Dialogue Skills with Internal and External Stakeholders
21 | I want to make the CEO's message resonate!
22 | Employee interviews: How much can I disclose?
23 | How to convert management's words into a "usable" form?
24 | Recruitment PR: Where should I start?
25 | How to restructure recruitment sites from a candidate's perspective
26 | Information leakage by employees... How to design guidelines
27 | How to do PR for cultural integration after a corporate merger?
28 | Anniversary projects: Communication techniques that don't end as "just a celebration"
29 | How to communicate public-private partnership projects?
30 | PR not trusted internally... Rebuilding internal branding

Part 2 | Adaptability and Practicality — How to Overcome High-Load PR
Chapter 4 | Acquiring "Communication Weapons" for the Field
31 | How to communicate a change in CEO? Design techniques for messages that "move people"
32 | Seminar presentation materials: What and how to say it to make an impact?
33 | How to get a CEO's contributed article published? From media selection to editorial support
34 | Book publishing as a PR strategy — From planning to generating buzz
35 | Novelty distribution copy: Words that move hearts without sounding like a sales pitch?
36 | Recruitment pitch materials: How to update to the latest "attraction points"
37 | Old company introduction materials: Where to start rebuilding?
38 | PR strategy to "win" awards and how to polish applications
39 | In-house PR events: PR techniques that balance attendance and media coverage
40 | Design skills to overcome the "who is the subject?" problem in collaboration projects

Chapter 5 | Turning Diverse Outputs into "Meaningful Communication"
41 | Research releases aren't hitting the mark... How to redesign for effective communication
42 | Interview articles are too long... How to find "words that convey everything in one phrase"
43 | Podcast delivery scenarios: AI utilization to polish into a "listenable structure"
44 | YouTube video descriptions: Designing a "flow that doesn't get skipped"
45 | Update the IR page! Maintenance techniques that balance readability and reliability
46 | FAQs and customer support copy: "Words that don't cause misunderstanding" polished with ChatGPT
47 | English copy for overseas distribution: How to translate cultural nuances?
48 | External positions/concurrent messages: How to express them with "consistent words"?
49 | Shareholder newsletters: Techniques for restructuring from an "investor perspective"
50 | PR for collaboration with local governments: How to bridge the language gap between government and private sector

Chapter 6 | Cultivating Brand "Character" through Words and Culture
51 | CSR x Regional Cooperation: PR that brings out "co-creation" rather than "support"
52 | ESG reports: From "unreadable materials" to "readable content"
53 | Internal newsletters on Slack... Posting techniques common to teams with high engagement
54 | Brand rebuilding phase: Inventorying words that fit the "current company"
55 | Alliance announcements: The secret to design that avoids arguments over "who is the subject?"
56 | Multilingual support: Why "cultural adaptation" is required instead of translation
57 | "Involvement-type" PR techniques to turn employee "vague dissatisfaction" into appeal
58 | Prompt design to verbalize company character through AI and employee workshops
59 | Mechanisms and "drawing out" techniques to have employees talk about their "favorite company"
60 | Advancing the "re-verbalization" of mission and vision while dialoguing with AI

Part 3 | Adaptability and Practicality — How to Overcome High-Load PR
Chapter 7 | What should PR do in times of crisis?
61 | How to avoid failure in announcing a CEO change — Organizing expectations, peace of mind, and the future
62 | New titles (like CPO): How to explain "titles no one has heard of"?
63 | Stock price plunge: Immediate judgment and preparation when media response can't keep up
64 | Messages during disasters: Copy design that balances "calmness and warmth"
65 | Employee health initiatives: Structure techniques that convey "feelings" rather than "systems"
66 | Communicating research results to the public: How to fuse "science" and "story"
67 | OSS project PR: Find common ground in language with engineers
68 | How to communicate patent acquisition? Techniques to convert "technical appeal" into PR
69 | Logo renewal: What is the one sentence that prevents internal confusion and conveys significance to the outside?
70 | Legal/governance reform: Utilizing AI for "writing that conveys without misunderstanding"

Chapter 8 | Fulfilling Accountability to Systems, Management, and Society
71 | Responding to legal changes: PR writing that doesn't end with "it wasn't read"
72 | Flooded with interviews and inquiries! Judgment criteria when media response can't keep up
73 | What is the PR perspective to turn board meeting materials into "readable materials"?
74 | Year-end reports: Redesigning to convey "gratitude" and "will"
75 | M&A announcements: To avoid misjudging "who the message is for"
76 | Industry-academia collaboration PR: Designing human touch that isn't conveyed by "results alone"
77 | Lobbying and PA activities: How to maximize "power of words" in cooperation with PR
78 | Consistency between IR and ESG: Techniques for re-editing as a corporate story
79 | Press conferences/announcements: What are the "scripts and lines" that don't leave the audience behind?
80 | 3 time-saving measures and AI utilization methods to use when "there is no time for PR"

Chapter 9 | Applications and Practices to Expand the Power of PR to Society
81 | I want to be on TV or the Nikkei newspaper! "How to find material" that works in the field
82 | Social issue-solving PR: How to turn "correctness" into "empathy"?
83 | SME PR: How to overcome the "only the president can talk" problem
84 | Freelance PR: How to create weapons to "get results even alone"
85 | How can local companies get into the media? What are the tricks to overcome "distance"?
86 | Retail, food/beverage, store business: "Street corner stories" to deliver to the media
87 | I tried working with a PR company, but... Points for "outsourcing without failure"
88 | PR services/AI tools: What are the "selection criteria" and "effective prompts"?
89 | Preventing trouble! Common PR accidents and their "prevention patterns"
90 | Community design and story construction to "increase company fans"

Part 4 | Future and Essence — Words Spun by AI and Humans
Chapter 10 | The Future of PR — The Intersection of Feelings and Technology
91 | Want to increase internal engagement? "Creating a culture of words" is the starting point
92 | How to co-create with influencers? Polishing the "bridging power" of words
93 | How do you verbalize corporate culture? PR perspective to cultivate "character"
94 | What is "company character"? Workshop design to bring out words with generative AI
95 | "Favorite company" project told by employees: How real should it be?
96 | AI and PR: What is the right distance? Searching for a relationship that is neither dependence nor rejection
97 | What should PR professionals leave to AI and what should they handle themselves?
98 | "Why do I do PR?" Continuing to ask this question adds depth to words
99 | What is the future of PR? How will it evolve amidst generative AI and social change?
100 | PR is a job that continues to connect "people's feelings"

01 | I can't write a press release! How should I convey this?

(Chapter 1 | Structure and Expression — Refining "Effective Design")

■ Introduction: "What should I write... time just passes while the page remains blank"

IT startup 'ConnectLink'.
With 80 employees, the only dedicated PR person is Nao Tamura.

Today's task is to 'create a press release for a new service launch'.

However, there is a blank Word file on the PC screen.
Only the cursor is blinking.

• 'I have the facts. But how do I write it so it becomes "news"?'

• 'If I write down all the service details, it will get too long...'

• 'I don't know what I should write first...'

Time is just ticking away.
The pressure from the office asking 'Is it done yet?' is also intense.

■ Case: 3-step process using AI for 'Information Organization -> Structure Design -> Draft Generation'

Nao first threw the following prompt into ChatGPT.

Prompt: 'Based on the service overview below, please create a "structure plan" for a press release aimed at the media. [Service Name]: [Features]: [Background]: [Target Media]: IT/Business [Tone]: Fact-based/Concise/Reliability-focused'

Next, she sent the structure plan and service description to Claude with the following prompt.

Prompt: 'Based on this content, please finish a formal first draft of a press release including a lead paragraph, body, and contact information. - Structure that is easy for media to treat as news - Include "social background" or "trends" in the introduction - If there are numbers, be sure to include them'

Furthermore, she researched 'recent similar release examples' using Perplexity to check the tone and expressions.

Finally, she used NotebookLM to check for any overlapping expressions with past company releases.

■ Solution: Created a template that allows for 'starting to write without hesitation'!

This release was completed following the flow below.

1. Fact organization -> Structure plan generation with ChatGPT

2. Body draft -> Generation with Claude

3. Expression tuning -> Case study research & reflection with Perplexity

4. Differentiation check from past releases -> Confirmation with NotebookLM

As a result, Tamura reduced the time required to create the first draft to less than half of what it used to be.
Moreover, it was well-received in internal reviews, with comments like, "It's become easier to understand!"

■ Summary of AI Tools and Prompts Used

Tool | Use Case | Prompt Used (Excerpt)
ChatGPT | Creating an outline | "Based on the following service information, please create a press release outline for the media."
Claude | Writing the body | "Based on the outline above, please create the first draft of an official press release."
Perplexity | Researching similar release examples | "Please tell me the stylistic trends of recent IT service press releases."
NotebookLM | Checking for overlap with past releases | "Please check if the following text has overlapping expressions with our past company releases."

■ Column: "Start with a 'Structural Diagram,' Not 'Text' for Releases"

What many people tend to do is "start writing the body text right away."

However, what PR needs is
"a 'blueprint' that allows the reader to understand the flow.".

AI can be your partner in that structural design.
First, consult with AI about the "flow," then add the words.

■ What should I do in these situations?

Q: I get asked, "What is newsworthy about this service?"...
→ You can organize it by asking ChatGPT, "Please extract three newsworthy points from a media perspective."

Q: Information comes in scattered from within the company and it's not coming together...
→ Feed everything into NotebookLM and have it categorized by "facts, background, and numerical data."

Q: I'm worried that the content overlaps with past releases.
→ A "duplicate expression check" prompt with Claude or NotebookLM is recommended.

■ Reader Workshop: Let's find the "next release topic" with AI

1. Have ChatGPT generate 10 candidates for internal topics that could become the next press release.

2. Search for "recent industry trends" on Perplexity to find the right timing.

3. Create 5 draft title patterns for the next release in advance using Claude.

■ Finally

The correct way to create a press release is
not to "write it all at once," but to "separate and organize it."

Structure → Body → Tone adjustment → Internal review → Final polish.
AI can be your partner throughout the entire process.

02 | Sending to media but getting zero replies... Reviewing selection and text

(Chapter 1 | Structure and Expression — Refining the "Communicative Design")

■ Introduction: "I sent it, but there's no response at all..."

Logistics tech company "LogiLink Solutions."
180 employees. The PR representative is veteran Tomoya Takahashi (35).

Recently, he had been struggling with a recurring problem.

• "I've sent press releases to many companies, but I never get a reply."

• "Zero inquiries from reporters. Even when I call, they just say 'we'll consider it' and that's it."

• "In the first place, is this media list even any good?"

Even though he was confident in the content of his releases, he had a strong feeling that
"they weren't reaching the media side."

■ Case: "Using AI to thoroughly review 'reporter perspectives' and 'recipients'"

Takahashi first conducted the following research using Perplexity.

Prompt: "Please tell me the major media outlets related to the logistics industry, the characteristics of the reporters in charge, and the themes they have covered in the past."

Next, he uploaded the company's existing media list (Excel file) and the past year's distribution results to NotebookLM.

He analyzed them with the following prompt.

Prompt: "Based on the following media list and past distribution history, please categorize the 'media that actually led to coverage' and 'unresponsive media.' Also, organize the 'types of media that are compatible with each release content trend' and propose a priority list for the next distribution."

Furthermore, he reviewed the text using ChatGPT with the following prompt.

Prompt: "Rewrite this press release email for the media into text that 'reporters would want to read.' - Clearly state at the beginning why you are sending it to this media - Summarize key points in bullet points - Clearly state the benefits of turning it into an article - Keep the tone concise and polite"

Finally, he generated a "media-specific distribution template" using Genspark and standardized it for future use.

■ Solution: "Text that resonates with each media outlet" was born

As a result, the following changes occurred in the next distribution.

• The priority of recipient media was clarified

• He was able to create distribution emails tailored to the different tones of each media outlet

• I actually received two media inquiries for the first time!

Takahashi realized, "Until now, I just settled for 'mass distribution to the whole company,' but just by using AI to 'select and refine,' the response is so different."

■ Summary of AI tools used and prompts

Tool | Use Case | Prompt (Excerpt)
Perplexity | Industry media list & reporter profile research | "Please tell me about major media outlets for the logistics industry and the characteristics of their reporters."
NotebookLM | Past distribution history analysis & media-specific performance classification | "Please classify the past distribution list into 'media with publication history' and 'unresponsive media'."
ChatGPT | Rewriting emails for media | "Please rewrite this email content to make it something reporters would want to read."
Genspark | Creating media-specific templates | "Please create distribution templates for IT media, logistics media, and business media, respectively."

■ Column: "Mass distribution no longer reaches anyone"

The media world is also busy.
With hundreds of releases arriving every day,
emails that don't clearly show "who, why, and why now" will get buried.

AI supports that "selective perspective" and the "effort of refining."Quality over quantity, design over speed.
This is
the "distribution power of next-generation PR."

■ What should I do in these situations?

Q: I feel like my media list is outdated to begin with...
→ Research the latest media and reporter lists with Perplexity, and cross-reference them with your own list using NotebookLM.

Q: My email drafts sound too much like templates.
→ Ask ChatGPT to "reflect the tone of articles recently published by this media outlet."

Q: My past distribution results are a mess in Excel...
→ Use the "Distribution Record Database Prompt" in NotebookLM to organize them before analyzing.

■ Reader Workshop: "Re-examine your company's media list with AI"

1. Upload your current media list & distribution history to NotebookLM

2. Create "distribution template drafts" with ChatGPT

3. Prepare a "collection of distribution templates by media type" with Genspark

4. List new media outlets with Perplexity and create an updated version

■ Conclusion

The era of "if you send it, they will read it" is over.

That is precisely why we should design "who, with what words, and how to deliver it" from scratch again, together with AI.

03 | Rebuilding "unreadable internal PR" through structure

(Chapter 1 | Structure and Expression - Refining 'Effective Design')

■ Introduction: 'I send out an internal newsletter every month, but...'

Pharmaceutical company 'Medical Force'.
600 employees, with branch offices scattered across the country.

Sayaka Akiyama,in her third year of PR,was struggling with the response to theinternal newslettershe published monthly.

• 'I feel like no one is reading it...'

• 'Even after sending it out, the "unopened rate" remains at 80%.'

• 'Even when I conduct surveys, the feedback is always "there were no articles that left a particular impression."'

The content includes event reports, comments from the president, and topics from each department...
But the feeling that 'it's not getting through' only continues to grow.

■ Case: 'Redesigning for a "Readable Structure" with AI'

First, Akiyama uploaded the past year's worth of internal newsletters toNotebookLM.

She objectively analyzed the issues using the following prompt.

Prompt: 'Analyze the following group of internal newsletter articles and point out three reasons why employees are skipping over them. Furthermore, please suggest improvements to the structure to make it an "internal newsletter that gets read." [Perspectives to include in improvement points]: - Flow design - Headline design - Distribution timing - Format improvements to increase completion rates'

The improvement suggestions that emerged were as follows.

• Reason 1: Too much information, leading to 'leaving before reading'

• Reason 2: Headlines make it difficult to tell 'who the information is for'

• Reason 3: President's comments are 'boilerplate' and lack impact

Next, she usedChatGPTto input the following prompt to strengthen headline ideas and flow design.

Prompt: 'To turn the following newsletter content into an "internal newsletter that gets read," please create the following: 1. Catchy top headline ideas (3 patterns) 2. Improvement examples for each article title 3. "Next action button wording" to increase completion rates'

Furthermore, she created a visually readable layout proposal inCanva. She also created an 'Internal Newsletter Improvement Prompt Collection' using
Genspark, leading to continuous improvements from the next issue onwards.

■ Solution: Transformed into an 'internal newsletter that people open and read!'

Starting from the next issue,

• Clearly state "Who is this information for?" at the beginning of each article

• Also add an impactful one-liner using ChatGPT to the CEO's message

• Placed a "Read this next!" button for each article, doubling the number of transitions on the internal viewing page

One week after distribution, the completion rate improved to 1.8 times higher than before.
Reactions from employees also increased, such as "I read the whole thing for the first time" and "I'm looking forward to the next one."

Akiyama feels the results, noting, "I've developed the habit of thinking about the 'reading design' using AI."

■ Summary of AI Tools and Prompts Used

Tool | Usage Scenario | Prompt Used (Excerpt)
NotebookLM | Analyzing past internal newsletter issues | "From the following newsletters, please identify the 'reasons for not being read' and 'improvement measures'."
ChatGPT | Headline, flow design, and tone improvement | "Please provide article title ideas and CTA copy that will make employees want to read."
Canva | Layout and visual improvement | "Please create an internal newsletter template that prioritizes readability."
Genspark | Creating a collection of improvement prompts for future issues | "Please create a collection of prompt templates for improving internal newsletters."

■ Column: "Don't 'make them read,' but 'make them want to read'"

A trap that PR professionals often fall into.

"This is important information, so they should read it"

But what is really needed is
"a design that reduces the reasons not to read and increases the reasons to read".

AI is a presence that can support that "shift in perspective to the reader's point of view"
calmly and quickly at any time.

■ What should I do in these situations?

Q: The CEO's message is repetitive...
→ Let's use ChatGPT to generate multiple patterns of "CEO message ideas that will make employees feel a little better tomorrow."

Q: The atmosphere from within the company is cold, like "Is it another internal newsletter..."
→ Let's use NotebookLM to analyze past issues and identify "article genres that had good reactions."

Q: The design is old-fashioned, and it's not being read because of how it looks
→ Let's create a "modern newsletter template" in Canva to give it a fresh look.

■ Reader Workshop: "Let's partially change the next internal newsletter using AI"

1. Analyze completion rates and reactions of past issues with NotebookLM

2. Generate title and catchphrase ideas for the next issue with ChatGPT

3. Revamping layouts with Canva

4. Customizing an 'Improvement Prompt Set' for your company using Genspark

■ In Conclusion

Internal communications is a
battle of design skills to eliminate the 'I thought I communicated it' gap.

If you use AI,
you can definitely create an 'internal newsletter people want to read.'

04 | How do you distinguish between IR materials and PR materials?

(Chapter 1 | Structure and Expression - Refining 'Communicative Design')

■ Introduction: 'Is this really okay for an IR document created by PR?'

Growth Lab, a startup preparing for an IPO.
Approximately 150 employees.

A request that suddenly landed on the desk of PR representative Takayuki Tanaka one day.

Tanaka, who was still vague about the difference between IR and PR, was bewildered by the
mountain of Excel and PowerPoint files sent from Corporate Planning.

• 'Isn't this too lacking in explanation to release externally as is?'

• 'But I shouldn't mess with the numbers on my own...'

• 'What is the difference between an 'IR document' and a 'PR document' anyway?'

There was no clear distinction even within the company.
The general sentiment was, 'Well, isn't it fine as long as PR makes it easier to understand?'
Tanaka was struggling all by himself.

■ Case: 'Visualizing the difference in roles with AI'

First, Tanaka uploaded the past three years of NotebookLM IR materials and press releases.

He analyzed the differences using the following prompt.

Prompt: 'Compare the following IR materials and press releases, and create a list of the differences in expression, structure, and tone between IR materials and PR materials. Also, create a checklist of points to keep in mind when re-editing IR materials from a PR perspective.'

The main differences identified are as follows.

Item IR Materials PR Materials Target Investors, analysts Media, general consumers Tone Number-based, objective Story-focused, readability-focused Level of detail Comprehensive numbers/metrics required Concise, focusing on key points Word choice Many financial terms General terms, plain language

Next, ChatGPT was used to input the following prompt.

Prompt: "Based on the content of this IR material, please restructure it for a 'media press release.' - Use language that is easy for general readers to understand - Keep numbers based on facts - Clearly supplement the growth story and background"

Furthermore, Claude was used to create **'examples of management comments'** with the following prompt.

Prompt: "Based on this content, please create two patterns of comments from the CEO: one for investors and one for the media."

■ Solution: 'The parts that PR needs to organize have become clear'

The 'IR x PR division of roles policy' summarized by Tanaka is as follows.

• IR materials rely on numbers, facts, and comprehensiveness → Corporate planning is the primary lead

• PR materials translate 'how this information impacts the world' → PR is the primary lead

• PR adjusts the tone for CEO comments and story sections

• Release text is created by re-reading IR information from a 'media perspective'

As a result, IR materials and PR releases were shared internally as separate finished products and the workflow for creating talk scripts for press conferences was also established.

■ Summary of AI tools and prompts used

Tool Usage Scenario Prompt Used (Excerpt)
NotebookLM Analyzing differences between IR and PR materials "Compare past IR materials and PR materials and list the differences in tone and structure"
ChatGPT Creating body text for PR releases "Restructure the content of this IR material into a release for general readers"
Claude Generating patterns for CEO comments "Create comment examples for investors and the media, respectively"

■ Column: 'Facts and readability can coexist'

There is no reason for PR to be weak with numbers.
However, the role of PR is to be a 'translator' that conveys those numbers.

• 'Do not change the facts'

• 'But, you can change how they are conveyed'

That perspective is the first step to 'breaking down the wall between IR and PR'.

AI can help you with that 'translation work' at high speed.

■ What should I do in these situations?

Q: The IR materials are full of technical jargon and I can't use them for PR...
→ Ask ChatGPT to 'translate this into language that even a middle school student can understand'.

Q: There's internal conflict over 'how much information we can disclose'
→ Use NotebookLM to list 'differences in tone from past disclosures' for an objective comparison.

Q: The CEO's comments are always templates and boring...
→ Have Claude create several 'future-oriented comment drafts that convey emotion'.

■ Reader Workshop: 'Translate IR materials once from a PR perspective'

1. Use NotebookLM to analyze the differences between past IR materials and press releases

2. Use ChatGPT to create a PR draft based on the latest IR materials

3. Use Claude to create alternative drafts for the CEO's comments

4. Finally, prepare for distribution through internal review

■ Conclusion

IR and PR are different, but they are connected.

'PR that conveys not just numbers, but also the background and the vision'.
That is the role of next-generation PR.

05 | How do you organize and communicate multiple services?

(Chapter 1 | Structure and Expression - Refining 'Effective Design')

■ Introduction: 'We have so many services that no one knows what we actually do...'

Marketing support company 'Axion Partners'.
10th year since establishment, 120 employees.

PR representative Shinji Kawaguchi has recently been struggling with a certain problem.

New businesses, subsidiary services, sub-brands...
Before he knew it, over 15 services of various sizes had sprung up within the company.

Sales materials, websites, press releases...
In every document, the state is such that 'there are so many service names listed that it's unclear what the main one is'.

• "I can't even answer when asked in one word, 'What kind of company is it?'"

• "Even with press releases, we argue every time about 'which service is the main focus?'"

• "Even from the media, it always ends with, 'You're a company that does a lot of different things, right?'..."

Kawaguchi was worried, thinking, "If we don't organize our services, both PR and sales will be lost..."

■ Case: "Visualizing the 'Overall Map' and 'Priorities' with AI"

First, Kawaguchi uploaded the company's service list, business division documents, and past press releases to NotebookLM.

He attempted to organize the information using the following prompt.

Prompt: "From the following service list and related documents, please create a 'Business Portfolio Map.' - Target customer segments for each service - Revenue contribution (to the extent known) - Priority for media information dissemination - Relationship map between services"

The output was the following organized result.

Service Category Target Revenue Impact PR Priority SaaS Marketing Tool B2B Marketing Manager High ◎ Consulting Service Corporate Planning/Business Development Dept Medium ○ Owned Media Operation Agency PR/Public Relations Staff Low △

Next, he entered the following prompt into Genspark.

Prompt: "Based on this portfolio map, please create a 'Service Introduction Order Guideline for PR Materials.' - Principle: Determine the order by multiplying revenue impact by PR priority - Each service description should be short, within 50 characters"

Furthermore, he created "one-sentence introduction drafts for each service category" using ChatGPT, and finally created a "Company Service Map at a Glance" visual using
Canva.

■ Solution: "The 'Master-Slave Relationship' between services was visualized"

In the new PR materials Kawaguchi created,

• The main service was clearly stated at the beginning

• Sub-services were concisely grouped

• The structure allows people to understand 'What does this company do?' in 3 seconds

As a result,

• The reception of press releases by reporters improved

• Well-received by staff, with comments that even sales materials have become 'easier to explain'

• Significantly reduced stress caused by information overload from a proliferation of services

■ Summary of AI Tools Used and Prompts

Tool | Use Case | Prompt (Excerpt)
NotebookLM | Structuring service information | "From the following materials, organize the target, revenue, and PR priority, and create a portfolio map."
Genspark | Creating sequence guidelines | "From this list of services, create a 'rule for the order of introduction in PR materials'."
ChatGPT | Creating one-sentence descriptions for each service | "Explain each service in 50 characters or less for a non-expert audience."
Canva | Visualizing service maps | "Create an infographic that shows the overall picture of company services at a glance."

■ Column: 'A company that doesn't get its message across is a company that isn't organized'

The more information a company has, the more it tends to succumb to the
'desire to say everything'.

However, the first thing PR should do is 'select, arrange, and trim'.

AI supports this process quickly and without hesitation.

What should be communicated first?
The answer always lies in data and organization.

■ What to do in these situations?

Q: Each department claims they are the 'main attraction'...
→ Use Genspark to create a 'priority decision criteria template for PR materials' and build consensus across the company in advance.

Q: The service descriptions are too long
→ Use the 'summarize to within 50 characters per service' prompt in ChatGPT to tidy them up.

Q: Sales materials and PR materials are inconsistent
→ Have NotebookLM read both and have the AI create a 'guideline for consistency in tone and order'.

■ Reader Workshop: 'Redesign the order of your company's service communication using AI'

1. Upload service lists and materials to NotebookLM

2. Create 'introduction order guidelines' with Genspark

3. Generate one-sentence description drafts with ChatGPT

4. Create a 'service map' visual with Canva

■ Conclusion

'Saying everything' is equivalent to
'communicating nothing.'

Select, narrow down, and arrange.
That is the 'compositional power' of public relations.

06 | No reaction to SNS posts... What is the true identity of a brand identity crisis?

(Chapter 1 | Composition and Expression - Refining 'Effective Design')

■ Introduction: 'Why aren't our posts getting any traction...?'

Apparel D2C brand 'Mellow Days'.
With 40 employees, PR and SNS management is handled bySatomi Yoshikawa (29), who holds a concurrent role.

X (formerly Twitter), Instagram, LinkedIn...
Even though she diligently posts to various platforms every day—

• 'The number of likes is low...'

• 'The number of followers isn't growing at all...'

• 'At first it was stylish, then it became serious, and recently it's been casual... the direction is all over the place.'

Yoshikawa suddenly realizes.

■ Case: 'Visualizing & Redesigning a Brand Identity Crisis with AI'

First, Yoshikawa uploads the past year's worth of SNS post data toNotebookLM.

She inputs the following prompt.

Prompt: 'From the following SNS post logs, please classify the patterns of post tone, style, and format. - Category examples: Friendly, Serious, Stylish, Explanatory, News-style, etc. - Please also provide the number of posts for each pattern and the average number of reactions (likes, comments, shares).'

The results the AI produced were as follows.

Tone Category | Number of Posts | Average Likes | Average Comments | Friendly Chat-style | 50 | 12 | 3 | Stylish Visual Posts | 30 | 28 | 6 | Serious Business Reports | 40 | 5 | 1

Next, she inputs the following prompt intoChatGPT.

Prompt: 'Based on the analysis of past SNS post data, please propose a recommended tone and manner policy for this account moving forward. - By platform - Assuming current follower demographics - Prioritizing post categories with high engagement'

Furthermore, she creates an 'SNS Tone and Manner Design Prompt Template Collection' usingGenspark.

Finally, we visualized the "design samples by tone and manner" using Canva and shared them with the team.

■ Solution: "Established a 'Core' for Post Tone"

The new policy Yoshikawa decided on is as follows.

• Instagram: Unified with a "Stylish & Lifestyle Tone"

• X: Emphasizing familiarity with "Casual Chat + Behind-the-Scenes Stories"

• LinkedIn: Focused on "Serious Business News"

As a result,

• The response rate per post improved to twice the previous level

• Started receiving DMs from followers saying, "I really like your posts lately!"

• Received positive feedback internally, such as "The tone has become consistent"

■ Summary of AI Tools Used and Prompts

Tool Usage Scenario Used Prompt (Excerpt)
NotebookLM Past SNS Post Analysis "Classify by post tone and calculate the average engagement." ChatGPT Tone and Manner Policy Proposal "Based on past post data from this SNS account, please provide a recommended future tone and manner policy." Genspark Tone and Manner Design Template Creation "Please create a collection of prompt templates for SNS tone and manner design."
Canva Tone and Manner Design Sample Creation "Please create SNS post sample designs for stylish, casual, and business news styles respectively."

■ Column: "Rather than 'Buzzworthy Words,' Choose 'Consistent Words'"

What is important in SNS operation is
not "searching for one-off buzzwords."

• "That's so like this company"

• "There's a sense of unity lately"

• "Oh, that tone again"

Being thought of in that way is
the first step toward "a company being loved on SNS".

AI helps you build that "core."

■ What should I do in these situations?

Q: I don't know the optimal tone for each platform.
→ Ask ChatGPT to provide 'proposed corporate tone options suitable for each platform'.

Q: To begin with, I don't know what the trends of our posts are...
→ Use NotebookLM to perform tone classification and engagement analysis on past posts.

Q: I want to change the design to match the tone as well.
→ Create 'tone-specific design templates' in Canva.

■ Reader Workshop: 'Let's visualize your company's current tone and manner using AI'

1. Load past post data into NotebookLM

2. Generate 3 patterns of 'optimal tone and manner policies' with ChatGPT

3. Create a 'tone and manner design prompt set' with Genspark

4. Create post design samples for each tone in Canva

■ Conclusion

The reason for a lack of response on social media is often
more about being 'tone-lost' than 'lacking content'.

Let's reclaim our 'authentic voice' together with AI

07 | Has your owned media stopped updating?

(Chapter 1 | Structure and Expression - Refining 'Effective Design')

■ Introduction: 'It's been "under preparation" for a year and hasn't moved...'

Human resource service company 'Human Career Solutions'.
Approximately 300 employees. PR team of 2.

PR Manager **Daisuke Nakano (38)** has been
assigned as the 'owned media manager' for a year now.

The internal project name is 'Brand Journal'.
Launch meetings, meetings with external production companies, brainstorming themes...
I've done a lot of preparation.

But.

• 'The first article hasn't been published yet...'

• 'I can't gather enough content'

• "I keep getting turned down when I ask for help with internal drafts."

• "I've been told to 'report on progress at the next meeting,' but nothing has moved forward..."

The owned media management dashboard had been sitting idle,
still filled with dummy 'test post' articles.

■ Case: "Using AI to streamline the 'idea generation to editorial workflow' all at once"

First, Nakano input the following prompt intoChatGPT.

Prompt: "Please list 30 article themes that are easy to handle on owned media in the human resources service industry. [Prerequisites]: - BtoB focus - Target readers are HR personnel and management - Wide range of topics including case studies, trends, know-how, and employee interviews."

Next, Nakano researched 'owned media examples from competitors' usingPerplexity.

Then, Nakano input the following prompt intoGenspark.

Prompt: "Based on the following list of 30 themes, please create the following output: 1. Priority publication order (high to low impact) 2. Draft titles for each article theme 3. A list of necessary departments and individuals to interview."

Furthermore, Nakano uploaded internal documents (past press releases, internal newsletters, CEO speech drafts, etc.) toNotebookLM.

The following prompt was used to extract candidates for repurposing past content.

Prompt: "From the following set of internal documents, please list 'information topics that could be repurposed for owned media.' - New service announcements - Employee interviews - Internal event reports - Past data that can be turned into columns."

Finally, Nakano created an 'article production progress management template' usingNotion AI and shared it with the team.

■ Solution: "Breaking through the 'can't publish the first article' syndrome!"

One month later, the following results were achieved.

• Thefirst commemorative article was finally published

• Themes and outlines for thenext three articleswere also finalized

• An 'Owned Media Committee' was officially launched within the company

• Nakano himself grasped the 'feeling of running an editorial workflow with AI'

■ Summary of AI tools and prompts used

Tool | Use Case | Usage Prompt (Excerpt)
ChatGPT | Article Ideation | "Give me 30 article ideas for an owned media site targeting the HR industry."
Perplexity | Competitor Case Research | "Tell me 5 successful owned media case studies in the HR services industry."
Genspark | Prioritization & Structure Design | "Rank these 30 ideas by impact, and provide draft titles and a list of internal interview candidates."
NotebookLM | Internal Content Mining | "From these internal documents, list topics that could be repurposed for our owned media."
Notion AI | Production Workflow Management | "Create a template for managing owned media article progress."

■ Column: "AI is effective for breaking through the '0 to 1' phase"

The biggest reason owned media stalls.
It is "not getting the first piece out.".

• Deciding themes

• Arranging interviews

• Creating outlines

• Drafting text

AI can handle all of this.
Humans only need to handle "direction decisions" and "final confirmation."

Start with one piece.
Everything starts moving from there.

■ What to do in these situations?

Q: "I don't know which department or who to interview."
→ Let's create an "internal interview target list" prompt in Genspark.

Q: "My writing is disorganized, and the structure is a mess."
→ Let's have ChatGPT generate an "800-character, readable outline based on this article topic."

Q: "I want to use past materials, but I don't know where anything is..."
→ Let's upload all past materials to NotebookLM and have the AI extract "repurposing ideas."

■ Reader Workshop: "Decide on your 'next piece' with AI"

1. Generate "30 article ideas for your owned media" using ChatGPT

2. Create a "priority list" and "draft title collection" using Genspark

3. Extract additional "internal topics" using NotebookLM

4. Create an "article progress chart" in Notion and start on the "first piece"

■ Finally

What a stalled media outlet needs is to "just publish something."

AI is a presence that gives you a push for that 'first step'.

08 | Designing for 'empathy' rather than aiming for 'buzz' on social media

(Chapter 1 | Structure and Expression - Refining 'Effective Design')

■ Introduction: 'It's not going viral... but I don't feel like aiming for viral is right either'

Educational SaaS company 'StudyLink'.
With 70 employees, social media management is handled solely by PR representative **Yu Nanjo (32)**.

Her recent worry is her 'non-viral social media account'.

• 'Competitors are getting tens of thousands of impressions every week...'

• 'Management keeps telling me to grow it more'

• 'Even when I try to post meme-like content to go viral, it just doesn't land'

Honestly... she was already exhausted.

Nanjo began searching for a way to create a 'social media presence that resonates' that felt true to them.

■ Case: 'Designing for "empathy" rather than "buzz" using AI'

First, Nanjo conducted the following research using Perplexity.

Prompt: 'Please provide 5 successful examples of "empathy-based posts" on social media for an educational SaaS company. Please prioritize examples that lead to "daily empathy and trust building" rather than "going viral."'

Furthermore, she uploaded the past year's worth of company social media post data to NotebookLM, and analyzed the reactions with the following prompt.

Prompt: 'From the following social media post logs, identify the post categories with the most "likes, comments, and saves." Additionally, summarize the "characteristics of posts that resonate" into 3 points.'

The answer derived by the AI was clear.

• Employee story-based content

• Customer success story interviews

• Relatable posts about 'everyday workplace life'

Based on this, I entered the following prompt into ChatGPT.

Prompt: 'Based on the following three categories, please generate 20 ideas for relatable social media posts. Categories: 1. Employee daily life and behind-the-scenes stories 2. Customer testimonials and case studies 3. Little relatable moments and struggles from the field'

Furthermore, I generated a 'collection of relatable social media post templates' using Genspark. I also created 'template designs for employee stories' using
Canva.

■ Solution: 'The start of a connected social media presence'

In subsequent social media operations,

• Although the growth in follower count was gradual, the 'engagement rate' increased significantly (about 2.5 times).
  

• Responses from customers in the comments section saying, 'This is the same for us!' increased.

• Direct messages started arriving saying, 'I like this account.'

• Positive feedback increased from within the company as well, such as 'The atmosphere of the social media account has changed' and 'It feels more natural and better than before.'

Nanjo realized once again that 'building grounded fans' is more important than aiming for viral hits.

■ Summary of AI tools used and prompts

Tool Usage Scenario Used Prompt (Excerpt)
Perplexity Competitor Case Research 'Please tell me 5 successful examples of relatable, rather than viral, social media posts.'
NotebookLM Past Post Analysis 'Please analyze the categories of posts with high engagement and their characteristics.'
ChatGPT Idea Generation 'Please generate 20 ideas for employee stories, customer case studies, and relatable posts.'
Genspark Template Creation 'Please create a collection of text templates for relatable social media posts.'
Canva Design Template Creation 'Please create social media post template designs for employee stories.'

■ Column: 'Rather than going viral, aim to leave an impression'

What social media managers often forget is
'lasting relationships,' not 'attention-grabbing posts'.

• You don't have to go viral every time.

• It's about someone thinking, 'Oh, it's this company again.'

• Today's post might lead to a business negotiation six months from now.

AI can properly assist with that kind of 'connection design' as well.

■ What should I do in these situations?

Q: "I always struggle with running out of ideas."
→ Use ChatGPT to generate a list of 50 empathetic social media post ideas.

Q: "My posts sound too much like I'm chasing viral trends."
→ Ask Claude to "rewrite this in a natural, down-to-earth tone."

Q: "I don't know which posts performed well."
→ Upload your past logs to NotebookLM and create a table ranked by engagement score.

■ Reader Workshop: "Let's design your next social media post for empathy with AI"

1. Analyze past engagement data with NotebookLM

2. Generate "next empathetic post ideas" with ChatGPT

3. Create draft post templates with Genspark

4. Polish post visuals with Canva

■ Conclusion

Social media is a
long conversation aimed at getting people to like you.

You don't need to go viral.
Let's build an account that people say they "love" with the help of AI.

09 | How much of the FAQ and CS copy should PR write?

(Chapter 1 | Structure and Expression - Refining "Communication Design")

■ Introduction: "More and more tasks are being dumped on PR..."

SaaS startup "Flow Systems."
90 employees. The PR representative is Misaki Ishikawa (31).

Recently, she had been receiving a series of requests from the CS (Customer Support) team.

• "Could you review the wording on the FAQ page from a PR perspective?"

• "People are saying the support email templates are too stiff..."

• "We want to align our user communication copy with the PR tone."

I'm happy to share information from a PR perspective, but
honestly... I'm at capacity.

PR, CS, Marketing—who does what and to what extent?
It felt like the "scope of responsibility for copy" was becoming increasingly ambiguous.

■ Case: "Organizing the 'Copy Domain Map' with AI"

First, Ishikawa uploaded the following to NotebookLM.

• Past press releases

• Support FAQ documents

• Email templates for customer inquiries

• Website FAQ page

Next, he visualized the "division of roles" using the following prompt.

Prompt: "Please classify the following documents by 'department responsible for information dissemination.' Classification category examples: PR / CS / Marketing / Product / Management. Additionally, please list the differences in 'writing style and tone.' Finally, please list 'points that need standardization' and 'areas where each department should consult with each other in the future.'"

The classification table generated by the AI showed the following results.

Content Type: Primary Responsibility, Current Tone, Points for PR to Monitor
Press Release: PR, Formal/Fact-based, Overall Supervision
Support FAQ: CS, Polite/Explanatory, Terminology Standardization/Tone Alignment
Reply Template: CS, Polite/Slightly Stiff, Support available as needed
Owned Media: PR, Story-driven, PR-led
Service Introduction Page: Marketing, Sales-oriented, Consult when necessary

Next, he used ChatGPT to output a "tone-standardized version of the FAQ copy."

Prompt: "Please refine the following FAQ text to be 'easy to read and friendly while aligning with the PR tone.' - Accurate to avoid misunderstandings - But with soft language - Length that won't tire the user out"

Furthermore, he created a "collection of prompts for standardizing CS copy tone" using Genspark. He also set up a "PR x CS collaboration copy check flow chart" using
Notion AI.

■ Solution: "The boundary of 'how much to do' became clear"

Ishikawa shared the following rules within the company.

• CS is primarily responsible for FAQs and support replies

• PR is responsible for "supervising tone guidelines" and "handling expression consultations"

• Hold a monthly "language alignment meeting" between PR and CS

• Automating tone consistency with AI as needed

As a result,

• The internal issue of "who looks at what" being unclear was resolved

• The clarity of FAQs was also improved

• A system was created where the PR workload does not become overwhelming

■ Summary of AI tools used and prompts

Tool | Use Case | Prompt (Excerpt)
NotebookLM | Categorizing current documents and organizing roles | "Classify each document into PR, CS, Marketing, Management, and Product, and list the differences in tone."
ChatGPT | FAQ tone adjustment | "Rewrite the FAQ text to be easy to understand from a user's perspective and consistent with the PR tone."
Genspark | Templating CS text | "Create a collection of prompt templates for standardizing the tone of CS communications."
Notion AI | Creating check flow templates | "Create a document review flow chart to be used by PR and CS."

■ Column: "You can create rules for 'who writes what and how much'"

The scope of PR writing gradually expands.

However, if you don't decide "how much to oversee" and "where to delegate,"
a "PR hell where you have to do everything" will arrive.

AI can also be used to create those boundaries.
Create rules, create templates, and reduce the burden.
That, too, is the work of next-generation PR.

■ What to do in these situations?

Q: "CS just thinks 'it's okay to just pass everything to PR'..."
→ Use Genspark to create a "Role Division Template for Internal Document Flow" and make it an internal rule.

Q: "The tone of the FAQs is inconsistent every time..."
→ Use ChatGPT to create a "Tone Standardization Prompt" and have AI proofread using the same criteria each time.

Q: "It's inefficient to create CS response text from scratch every time"
→ Use NotebookLM to organize past reply data and have AI generate "templates by common question pattern."

■ Reader Workshop: "Let's create an 'Internal Document Flow Chart' with AI"

1. Analyze your company's document stock with NotebookLM

2. Create "document templates usable by PR, CS, and Marketing" with ChatGPT

3. Create a "Role Division Manual" with Genspark

4. Share the flow chart in Notion and announce it company-wide

■ In Conclusion

Public relations is not about 'doing everything'.

To be the
'guardian of language quality,' let's work with AI to draw the line on how much to get involved.

10 | Techniques for Rediscovering Press Release 'Topics' When You've Run Out

(Chapter 1 | Structure and Expression - Refining 'Effective Design')

■ Introduction: 'I'm out of ideas...'

IT solutions company 'BizTech Link.'
With 250 employees, the dedicated PR person is Hayato Nakahori (37).

In the last six months, he has already issued 7 press releases.

• 'New Service Launch'

• 'Fundraising'

• 'Alliance Announcement'

But now, he has... no more topics.

Within the company, voices asking 'When is the next release coming out?' are only increasing.
Meetings to find topics keep coming up empty.
Nakahori sighed while opening the folder of past releases in the archives.

■ Case: 'Digging up "buried news seeds" with AI'

Nakahori first uploaded the following internal information from the past three years to NotebookLM.

• Press releases

• Internal newsletters

• Internal project documents

• Employee recognition records

• List of new customer implementation cases

Then, input the following prompt.

Prompt: "From the following set of internal information, list up 'news candidates that have not been released yet.' Extraction criteria: - Items with potential for external communication - Items with novelty and uniqueness - Items that could be turned into employee or customer stories - Internal projects likely to have developments in the next few months"

The output candidates are as follows.

• 'Internal growth topic: Exceeding 200 employees'

• 'Story of the birth of a female leader in the development department'

• 'Large-scale implementation case for a mid-sized client company'

• 'Internal sustainability initiatives'

Next, use ChatGPT with the following prompt to generate release ideas.

Prompt: "From the following list of news candidates, pick 3 'themes that could be issued as external press releases,' and propose a title idea, lead sentence example, and appeal points for the media for each."

Furthermore, use Perplexity to research recent release trends of competitors. Confirm 'whether similar cases have already been released' and 'what angles are new.'

Finally, use Genspark to create a 'prompt template collection for future topic discovery' and standardize future operational flows.

■ Solution: 'What is taken for granted internally is actually news'

The next release theme finally decided by Nakahori is,

• 'Behind the scenes of exceeding 200 employees & work-style reform'

Moreover, it was designed to also incorporate the story of the female leader's emergence.

After the release was distributed,
an interview offer for a 'Fast-Growing Company Feature' came in from an industry media outlet,
allowing us to overcome the 'period with no topics.'

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Used Prompt (Excerpt)
NotebookLM | Extracting 'buried topics' from internal information | 'Please list up internal news candidates that have not been announced in the past'
ChatGPT | Creating release ideas by theme | 'Please provide 3 ideas for titles, leads, and media appeal points for this theme'
Perplexity | Researching competitor cases | 'Please tell me the press release trends in the IT industry for the last 3 months' | Genspark | Creating templates for future topic discovery | 'Please create a prompt template collection for internal topic discovery'

■ Column: 'It's not that there is "nothing." It's just that you can't "see it."'

The problem of having no ideas for press releases.
Actually, it might just be that you lack the habit of looking at your company from an 'external perspective'.

• Changes in the number of employees

• Internal company events

• Small initiatives

• Customer feedback

Everything can become 'PR material' if you put it into words.

AI can help you find those 'buried seeds'.

■ What should you do in these situations?

Q: 'I always worry about whether this will actually be news.'
-> Ask ChatGPT to provide 'angles from a media perspective' for this information.

Q: 'Internal information is too scattered, making it hard to find.'
-> Upload everything to NotebookLM and organize it all at once using an 'unpublished material extraction prompt'.

Q: 'I'm afraid I'll run out of ideas again next time...'
-> It is recommended to create a 'monthly idea generation template collection' using Genspark.

■ Reader Workshop: 'Let's create a "Break the Idea Drought List" using AI'

1. Consolidate past internal data with NotebookLM

2. Generate '10 candidate themes for the next release' with ChatGPT

3. Check industry release trends with Perplexity

4. Create an 'idea discovery worksheet template' with Genspark

■ Finally

Ideas are not something that just 'fall from the sky'.

'They are something you have to go dig for yourself'.

AI can serve as both your shovel and a 'gold vein detector' for the soil you've dug up.

11 | PR policy is not set! Initial strategy design

(Chapter 2 | Verbalizing Strategy - Creating the 'Core' of PR)

■ Introduction: 'Even if I'm told to get more exposure...'

Mid-sized IT company 'B Innovations'.
About 300 employees, 15th year since establishment.

Recently, there was an organizational restructuring and service integration, and the PR team now has 3 members.

However, these voices were flying around within the team.

• 'Management tells me to increase exposure, but how should I act?'

• 'Recruitment PR, PR, IR... there are mountains of things to do, but the priorities aren't decided'

• 'Even when trying to decide on KPIs, the "purpose of doing it" isn't organized...'

Monthly reports list numbers like "number of articles published," "SNS follower growth," and "number of press releases," but
the whole team kept working while feeling uneasy, wondering, "What is this activity leading to?".

■ Case: "Verbalizing the 'Core of PR' with AI"

First, PR team leader Eri Ando uploaded the following internal documents to NotebookLM.

• Past PR reports

• Recent management planning documents

• Goal documents for each business division

• Past instruction memos from management

Next, input the following prompt.

Prompt: "From the following document group, read 'what management expects from PR' and output the following: 1. Hypothesis of goals expected of PR (up to 3) 2. Mapping of how current PR activities contribute to which goals 3. Point out any missing strategic areas"

Next, execute the following prompt in ChatGPT.

Prompt: "Based on the following management documents and PR report content, create a 'draft PR policy for the next year'. Conditions: - Clearly state mission, target, key themes, and main KPIs - Format that is easy for management to review - Volume within 800 characters"

Furthermore, Genspark was used to create a "Public Relations Policy Formulation Prompt Template Collection," organized so it can be used for future policy updates.

■ Solution: "We created a document that we can confidently call our PR policy"

Ando and the team finalized the policy through the following steps.

1. Visualize connection points with the management plan

2. Categorize current PR activities into three axes: "Recruitment Support," "Customer Awareness," and "Business Understanding Promotion"

3. Share with the entire company as the "FY2025 PR Policy Document (1-page A4 version)"

Even at the management meeting,

• "Now we can see what PR is doing."

• "With this policy, it's easier to collaborate with business departments."

there was a positive reaction.

The PR team finally gained a "guiding axis" to keep them on track.

■ Summary of AI Tools and Prompts Used

Tool | Use Case | Prompt (Excerpt) | NotebookLM | Goal analysis of internal documents | "From this set of documents, please hypothesize three things that management expects from PR." | ChatGPT | Drafting PR policy | "Summarize a one-year PR policy within 800 characters, including mission, target, theme, and KPIs." | Genspark | Creating policy formulation templates for future use | "Please create a collection of PR policy formulation prompt templates."

■ Column: "PR without a policy is like being lost"

The busier the PR team is, the more they tend to be
swept away by immediate tasks.

However, by using AI to organize past documents and articulating "Why are we doing this PR activity now?", everyone's approach changes.

Because there is a policy,
you can also choose "what not to do."

■ What should I do in these situations?

Q: "Management only tells me to 'just increase exposure.'"
→ Use ChatGPT to generate a list of potential PR goals, and start by discussing three options with management.

Q: "Currently, our PR activities are scattered and I can't organize them."
→ Have NotebookLM read your past reports and create an "Activity Mapping Table."

Q: "I'm worried that even if I create a policy, it will be back to square one in a year."
→ Let's use Genspark to create a 'Public Relations Policy Update Prompt Collection' to establish a system that can be updated every year.

■ Reader Workshop: "Create Your Company's 'One-Page PR Policy Sheet' with AI"

1. Analyze internal documents (management plans, PR track records) with NotebookLM

2. Create a policy draft with ChatGPT

3. Save the 'Policy Formulation Prompt Collection' in Genspark

4. Create a company-wide PR policy page in Notion

■ Conclusion

Public relations is a job that aims for "being meaningful" rather than "being seen".

"Why is PR doing this work this year?"

Let's put that answer into words together with AI.

12 | "Are PR KPIs necessary?" Breaking away from vague evaluations

(Chapter 2 | Verbalizing Strategy - Creating the 'Core' of PR)

■ Introduction: "I'm working hard, but I'm not being evaluated..."

BtoB SaaS company 'CloudFlow'.
About 200 employees, PR representative is Shota Matsui (33).

This past year, I've worked hard on press releases, events, and social media.

But one comment made at the management meeting stuck in my heart.

• "The number of articles published is increasing!"

• "Social media followers have also exceeded 1,000!"

Even when I said that, the reaction was lukewarm.

I'm not being evaluated. The budget isn't increasing either.
"What even are PR KPIs..."
Matsui sighed in front of his PC.

■ Case: "Taking on the challenge of 'storytelling with KPIs' using AI"

First, Matsui uploaded the following documents to NotebookLM.

• PR activity reports for the past year

• Press release distribution records

• Trends in media coverage volume

• SNS analysis reports

• Data from other departments such as hiring numbers and sales material downloads

Next, he used the following prompt to gain an objective view of the current situation.

Prompt: "From the following set of documents, create a list of PR activity volume, performance indicators, and points of change over the past year. Furthermore, provide three hypothesis-based patterns on 'how current KPIs connect to management issues.'"

Next, he used the following prompt in ChatGPT.

Prompt: "Based on the following PR performance data and company goals, create a 'draft PR KPI proposal for management meetings.' Required items: - PR mission - KPI candidates (categorized into volume, quality, action, and outcome indicators) - Hypotheses on connections to management indicators - Future improvement measures"

Furthermore, he created a 'PR KPI design prompt template collection' using Genspark.

Finally, he researched 'PR KPI examples from other companies' using Perplexity and incorporated them into the internal presentation materials.

■ Solution: "For the first time, I was able to articulate 'convincing KPIs'"

The 'PR KPI draft' that Matsui submitted turned out like this.

Indicator Category: KPI Example Volume
Number of press releases, number of SNS posts: Media contact frequency = PR activity volume Quality
Media coverage quality score: SNS engagement rate = PR output quality Action
Web traffic via articles: Number of recruitment entries triggered = Triggering internal/external actions Outcome
Increase in sales leads, increase in hiring: Changes in brand indicators = Contribution to business results

Furthermore,

• "I've been told I'm 'PR without numbers' until now, but if I organize it like this, I can explain it!"

• Let's set up a quarterly KPI review meeting from now on.

With that, a positive atmosphere emerged within the PR team.

■ Summary of AI Tools Used and Prompts

Tool | Use Case | Prompt (Excerpt)
NotebookLM | Analysis of past PR achievements | "List PR activity volume, results, and changes, and hypothesize their relationship with management indicators."
ChatGPT | Creation of KPI proposal materials | "Create PR KPI candidates categorized into volume, quality, action, and results, and show their relevance to management issues."
Genspark | Creation of KPI design templates | "Create a collection of prompt templates for PR KPI design."
Perplexity | Research on other companies' cases | "Tell me about PR KPI examples for 5 SaaS companies."

■ Column: "PR that cannot be measured is PR that will not be evaluated"

"PR results are hard to see."
"They cannot be quantified."

—Is that really true?

There are many things that can be visualized.

Moreover, with AI, you can objectively organize them from the perspectives of
"volume," "quality," "action," and "results."

If you want to be evaluated, start with
"the ability to translate things into measurable terms."

■ What to do in these situations?

Q: "We haven't been tracking numbers until now..."
→ Have NotebookLM read your past reports and generate "KPI proposals that can be created with the numbers you currently have."

Q: "I don't know what kind of indicators to set."
→ Have ChatGPT generate "examples of PR KPIs for B2B companies" and customize them for your own company.

Q: "I'm worried because many of my activities don't directly lead to results."
→ Use Genspark to create a "KPI classification template for each activity" and start by "classifying" them.

■ Reader Workshop: "Let's create KPI proposals with AI using 'existing data'"

1. Analyze PR performance data in bulk with NotebookLM

2. Generate a "PR KPI draft proposal" with ChatGPT

3. Create a "KPI proposal material template for the next meeting" with Genspark

4. Research and reference KPI examples from competitors using Perplexity

■ Conclusion

The era when PR was called an "unappreciated profession"
should come to an end.

“Demonstrate your value through words, numbers, and strategy.”

AI is the best partner for that “visualization work.”

13 | The Intersection of PR and IR: Information Design for Discussing the Future

(Chapter 2 | Verbalizing Strategy — Building the “Core” of PR)

■ Introduction: “What do you mean by ‘add a PR touch to the IR materials’?”

SaaS venture “GrowthLinks.”
120 employees, preparing for an IPO.

PR representative Takuya Omori (34) was suddenly told this by the Corporate Planning Department.

Numbers, numbers, and more numbers...
IR materials tend to become just a list of facts.

• “Where am I supposed to add a ‘PR touch’?”

• “If I add a narrative in a weird way, I’m worried investors might lose trust...”

• “But, ‘make it an IR material that shows the future’... how do I do that?”

Omori was struggling with the question, “Where is the intersection between PR and IR?”

■ Case: “Designing ‘IR materials that speak to the future’ with AI”

First, Omori uploaded the following to NotebookLM.

• IR presentation materials from the last two years

• PR releases (related to service strategy and business vision)

• Past statements and speech notes from the CEO

• Medium-term management plan documents

Input the following prompt.

Prompt: "Compare the following IR materials and PR materials, and list the 'elements of future perspective and business story missing from the IR materials.' Also, propose three improvement points to create an 'information design that makes it easy for investors to envision the future.'"

Next, create a 'draft lead paragraph for IR materials with a future perspective' using ChatGPT.

Prompt: "Based on the content of these IR materials, please create a 'lead paragraph that allows investors to grasp the outline of future strategy.' Conditions: - Keep numbers accurate - Concisely show the growth story toward the future - Tone should be sincere and positive"

Furthermore, create a 'draft CEO comment' using Claude with the following prompt.

Prompt: "Based on this content, please create two versions of a 'future-oriented CEO comment': one for IR and one for the media.
- For IR: Investor perspective, numbers, and growth potential
- For Media: Impact on society and storytelling"

Finally, create a 'collection of story design prompts for IR materials' using Genspark.

■ Solution: 'The IR materials have transformed into materials that tell a story of the future'

The completed materials are,

• solid with numbers as before

• but structured to clearly convey 'why this business is worth investing in'

• The opening lead concisely organizes market background -> company strengths -> future challenges

• and the CEO comment is now a dual-purpose piece that resonates with both investors and the media

It was highly rated by the internal IR team as well,

• 'It feels more impactful than usual'

• 'This is easy to use for analyst briefings'

they said.

■ Summary of AI tools and prompts used

Tool Usage Scenario Used Prompt (Excerpt)
NotebookLM Analysis of missing elements in current IR materials 'Compare IR materials and PR materials, and point out missing elements of future perspective'
ChatGPT Creation of future-oriented lead paragraph 'Create an IR lead paragraph draft that allows investors to grasp the future vision'
Claude Creation of CEO comments 'Create draft CEO comments for both IR and media'
Genspark Creation of story design templates 'Create a collection of story design prompts that can be used for IR materials'

■ Column: 'The sense of expectation for the future that lies beyond the facts'

The reason IR materials look cold is because they
only end up stacking facts.

However, with AI,

• Numbers → Background → Strategy → Future

You can design the story of this “flow of information”
in a short amount of time.

“IR that shows the future”
—that is an important area where PR can get involved.

■ What should you do in these situations?

Q: “I’m worried about adding ‘storytelling’ to IR materials.”
→ Use ChatGPT to generate ‘examples of investor-oriented story expressions’ and get creative while keeping risks low.

Q: “Management tells me ‘IR materials should only contain facts.’”
→ Use NotebookLM to compare ‘future-oriented sections of other companies’ IR materials’ and create materials for internal explanation.

Q: “The CEO’s comments are always formal and don’t resonate…”
→ Have Claude generate ‘drafts of future-oriented comments that excite investors.’

■ Reader Workshop: “Create a template for ‘IR materials that show the future’ using AI”

1. Use NotebookLM to compare and analyze past IR materials and PR materials

2. Use ChatGPT to create a ‘future-oriented lead sentence draft’

3. Use Claude to create multiple CEO comment drafts

4. Use Genspark to create an ‘IR story design prompt collection’ and utilize it for future standardization

■ Finally

When you stand at the intersection of IR and PR,
there are things that become visible.

That is, “the company’s future that lies beyond the numbers.”.

AI is your guide on that path.
Gain the power to become a ‘PR professional who can talk about the future’ right here, right now.

14 | The Big Picture of IPO Preparation PR and ‘Timing of Words’

(Chapter 2 | Verbalizing Strategy — Building the ‘Core’ of PR)

■ Introduction: From the day you hear "IPO preparation," PR time accelerates

Mid-sized SaaS company, Company S. Established 9 years ago, with 220 employees.
After Series D, they have finally entered the IPO preparation phase.

One morning, the PR manager, Mai Sugiyama, is told this after a management meeting.

For a moment, her heart skipped a beat.

IR shouldn't be her responsibility yet.
But where should she start with "messaging conscious of an IPO"?
When, what, and how much can be disclosed—

There is no clear roadmap within the company, and the PR department remains unclear about "how much they should be involved."

• "At what timing should we talk about our 'future vision'?"

• "How do we separate recruitment and culture messaging from IR?"

• "When does the 'quiet period' begin?"

While hesitant, Sugiyama began to draw up a "PR blueprint for the IPO preparation period" with the help of generative AI.

■ Case: The "5 Questions" that should be designed first for IPO PR

The questions Sugiyama faced were as follows.

1. When should PR get involved in IPO preparation?

2. How should the division of roles between PR and IR be organized?

3. What should be communicated in each phase of messaging?

4. How should we prepare for anticipated questions and risk information?

5. How can we deliver a "common story" to both internal and external audiences?

To these questions without clear answers, Sugiyama used ChatGPT, Claude, NotebookLM, and Perplexity to gradually refine the words and structure.

■ Solution: Visualizing the "PR blueprint for the IPO preparation phase" with generative AI

1. Designing a "PR Roadmap for IPO Preparation" with ChatGPT

Prompt: "Organize the overall picture of PR activities during the IPO preparation period in 'chronological order.' Target: For PR managers. Please design it in 'three phases,' including objectives, necessary communications, points to note, and coordination with related departments."

→ A step-based visualization divided into '1. Preparation Phase,' '2. Pre-listing Phase,' and '3. Post-listing Phase' was completed. This provides a foundation that is easy to explain within the company.

2. Refining the "Tone of Voice Suitable for IPO PR" with Claude

Prompt: "As a tone for PR statements issued during IPO preparation, it is necessary to balance 'consideration to avoid misunderstanding' with 'a hopeful future.' Please adjust the tone, nuance, and reader perception for the following three examples."

→ It was refined into a balanced tone suitable for pre-listing, such as 'modest but forward-looking' and 'accurate but approachable.'

3. Analyzing the "Consistency of IR, PR, and Recruitment PR" with NotebookLM

Sugiyama uploaded the following materials to NotebookLM:

• Mid-term Management Plan (IR material)

• Recruitment Branding Materials

• Mission and Value Content created by the PR Department

Prompt: "Compare the differences in 'expressions of future direction' and 'key messages' in these three documents, and provide paraphrasing suggestions to create a sense of unity."

→ Variations in tone and vocabulary were organized, leading to a state where the same vision can be articulated by anyone, anywhere.

4. Researching "Other Companies' IPO PR Cases and Precautions" with Perplexity

Prompt: "Research PR cases of IT companies that have had an IPO in the last three years and tell me about 'failed cases and the reasons for them.' Summarize the points to be careful about when communicating before listing."

→ By understanding other companies' 'flaming cases' and 'timing mistakes in communication,' I gained a decision-making axis to avoid landmines.

■ Summary of AI Tools Used and Prompts

Tool | Usage Scenario | Prompt Used (Excerpt) | ChatGPT | PR design for each IPO preparation phase | "Design PR activities for IPO preparation in 'three phases'" | Claude | Writing style and tone suitable for IPO PR | "Adjust to a tone that balances 'hope and calmness'" | NotebookLM | Unifying the vision for IR, recruitment, and PR | "Visualize the differences in 'how the future is described' across the 3 documents and provide paraphrasing suggestions for standardization" | Perplexity | Cross-sectional survey of other companies' IPO PR cases | "Tell me about 'PR trouble cases' of recent IPO companies. Include why they happened."

■ Column: What is required for IPO PR is "Quiet Passion"

Communication before an IPO requires "quiet passion" rather than flashiness. Excessive promotion creates misunderstandings, while remaining silent causes you to miss opportunities.

What is important is the ability to distinguish between 'what should be said now' and 'what should not yet be said'.
Generative AI serves as a guide to help refine this 'maturation and timing of information'.

■ What to do in these situations?

Q: I don't know when I should announce our vision.
→ Let's use ChatGPT to organize the 'communication phases during IPO preparation'.

Q: The messages from IR, PR, and recruitment are inconsistent.
→ You can use NotebookLM to compare and unify the key messages of the three documents.

Q: Management is pushing me to 'be more aggressive'.
→ Use Claude to adjust the tone to be both cautious and ambitious. Checking examples of other companies' failures on Perplexity is also effective.

■ Reader Workshop: 3 steps to create a 'word plan' for the IPO phase

1. Design an 'IPO PR Roadmap' with ChatGPT

2. Refine 'tone and consideration' with Claude

3. Verify 'message consistency with other departments' using NotebookLM

■ Conclusion: Going public means opening up 'words of trust' to society

PR during the IPO preparation period is not just about making announcements.
It is an act of 'sharing' the company's future vision with society, and it is a moment where the company is tested on whether it is a 'trusted entity'.

Generative AI becomes a guide to turning that trust into words.

You, too, can surely become someone who can talk about the 'future of a company' through the power of PR.

15 | How to communicate governance? Techniques for expressing transparency

(Chapter 2 | Verbalizing Strategy - Creating the 'Core' of PR)

■ Introduction: The Audit and Supervisory Committee member refused an interview!? Who decided that 'governance cannot be talked about'?

Specialized Trading Company F. Although it is a long-established company, it has been focusing on strengthening ESG and governance for the past few years.
In recent years, they have also been working on communicating in non-financial areas, and from a PR standpoint, they wanted to actively convey information about 'governance, ethics, and transparency'.

However, PR Manager Masahiro Kawada hits a wall.

Cautious opinions emerged one after another from the internal management department, legal department, and IR.
Furthermore, there was no reply from the Audit and Supervisory Committee member to whom the interview request was sent.

Kawada opened Generative AI alone and set out on a journey to put into words things that had previously been inexpressible.

■ Case: 'Governance = Internal Organizational Chart' fails to communicate

Communication regarding governance at many companies is in the following state:

• It ends with just explanations of systems like 'Board of Directors composition' and 'Audit & Supervisory Board system'

• It leaves out 'why' that structure exists and 'how' decisions are made

• Fearing legal review, the tone becomes dry and results in 'documents that no one reads'

What Kawada aimed for was to communicate 'governance as a mindset,' rather than as a system, in a structure that resonates with the reader.

■ Solution: Redesigning 'transparency' from the structure up using Generative AI

1. Converting 'abstract values' into 'explainable language' with Claude

Prompt: "Please rewrite 'corporate transparency,' 'management soundness,' and 'convincing governance structure' into language that even the general public can understand. For example, please express 'a high ratio of external directors' as 'incorporating external perspectives on a daily basis.'"

→ Complex structural explanations were converted into a 'context of conviction,' creating a structure that 'answers the reader's questions'.

2. Comparing the consistency of 'internal system documents and philosophy documents' with NotebookLM

Kawada uploaded the following three types of internal documents:

• Corporate Governance Report (formal document)

• Management philosophy and company motto documents

• Past messages from the representative

Prompt: "Compare these three documents and analyze whether 'what the company values' is consistently communicated. Please specify any discrepancies in expression or tone, or divergence in keywords."

→ While they matched on the surface, it became clear that there was a subtle discrepancy between the 'explanation of external evaluation' and the 'words of management.' This became the basis for a trustworthy re-edit.

3. Generating 'internal and external explanation scripts' with ChatGPT

Prompt: "Please create messages explaining 'why our governance structure is the way it is' for the following three audiences: 1. Internal employees, 2. Consumers, and 3. Investors. Avoid jargon and convey the purpose, background, and underlying philosophy."

→ Enables 'tailored communication' for each target audience. Messages that can also be repurposed for company-wide meetings and recruitment PR are now ready.

④ Create a 'convincing infographic' using 'Structure + Message' in Canva

Prompt: "Please propose a diagram that can visually explain our governance structure using the following composition: 1. Management composition, 2. Decision-making flow, 3. Division of roles between external perspectives and internal checks, 4. An answer to the reader's question of 'why this structure?'"

→ Built new materials that deliver a 'structure that communicates through diagrams alone' paired with 'convincing explanations.'

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Prompt (Excerpt)
Claude | Simple paraphrasing of governance concepts | "Rewrite 'management integrity' in words that consumers can understand."
NotebookLM | Checking discrepancies between systems and philosophy | "Analyze whether there is 'consistency' between the governance report and the corporate philosophy."
ChatGPT | Creating scripts for internal and external audiences | "Create 'explanation patterns' for the governance structure by target audience."
Canva | Designing infographic composition | "Propose a 4-item diagram structure that conveys the meaning of the system."

■ Column: From superficial transparency to 'thoughtful transparency'

The word 'transparency' does not mean 'disclosing everything,' but rather 'communicating in a way that is convincing.'

Simply listing numbers and systems does not create trust.
Management philosophy, the background of decisions, and mechanisms for external checks—
The structure that conveys the 'why' behind these is the essence of true governance communication.

Generative AI is a 'partner that helps you rediscover the values your company cherishes' in the form of words.

■ What to do in these situations?

Q: People in the company say 'governance is the job of Legal and IR'
→ Use ChatGPT to organize the 'significance of speaking to consumers' and demonstrate your role as PR.

Q: The language used for philosophy and system explanations is inconsistent
→ Use NotebookLM to compare multiple documents, which allows you to objectively see the discrepancies.

Q: The auditor won't talk. I can't get information out...
→ Use Claude to construct 'sentences that give the system explanation a human touch,' and propose it as a 'communication script' rather than an interview.

■ Reader Workshop: 3 steps to turn governance into a 'readable story'

1. Convert 'System → Empathy Expression' using Claude

2. Check consistency between 'Philosophy and Structure' using NotebookLM

3. Prepare 'Explanatory text for each stakeholder' using ChatGPT

■ Finally: 'I can trust this company'—the source of that feeling is governance

A trusted company is one that shares the 'basis for its decisions'.
The messages it sends out contain the 'why'.

It is not just about the system, but talking about the reasons why the system was created and the reality of its operation that is the 'governance communication' that only PR can do.

16 | How to verbalize a company's 'reason for being'

(Chapter 2 | Verbalizing Strategy - Creating the 'Core' of PR)

■ Introduction: When asked, 'Who is the mission for?'

Healthcare venture, Wellsense Inc.. Five years since its founding, it is growing rapidly in the preventive medicine sector. Although it has 50 employees, it has been increasing its presence in the industry.

One day, while preparing for a presentation at a conference, the speaker and representative, Ihara, muttered.

The PR manager who was told this, Shizuka Akiyama, panicked inwardly.
It is true that 'Mission: Supporting life.' is written in large letters on the company bulletin board and recruitment page.

But—

• It is vague what it specifically refers to

• The connection to service content and business growth is hard to see

• It remains just words that make it look 'like social contribution'

Akiyama once again began the challenge of putting the company's 'reason for being' into words with generative AI.

■ Case: Has the mission become 'words floating in the air'?

What Akiyama was facing was the following challenge that many startups have.

• The company's mission or purpose is too abstract for anyone to talk about

• Recruitment candidates ask, 'You say nice things, but what kind of company is it?'

• At the beginning of presentations and press releases, the mission is treated like a 'decoration'

So Akiyama began to utilize generative AI as a 'partner for unearthing words'.

■ Solution: Unearthing and re-cultivating 'raison d'etre' with generative AI

① Designing questions with ChatGPT to organize the 'intersection of the company's origins and its future'

Prompt: 'I want to delve into what our mission, "Supporting Life," specifically means. Please create questions from the following perspectives: - Background of the founding - Connection to current business - What social changes we will respond to in the future - The "sense of purpose" and "pride" felt by employees within the company -> Please also include auxiliary questions for each that make it easy to start talking.'

→ Visualizing the "questions to consider before verbalization." Designed to be usable for interviews and internal dialogues as well.

② Outputting multiple 'raison d'etre cards' with Genspark

Prompt: 'Based on the mission "Supporting Life," please create 10 patterns of text cards that explain the "company's raison d'etre" from different angles. Please structure them by target, such as: 1. For citizens, 2. For medical professionals, 3. For government/local authorities, 4. For investors, 5. For employees, etc.'

→ Words that resonate with both the public and business sectors are gathered, and a map of "who to talk to" and "how to talk to them" is completed.

③ Analyzing 'past PR materials x founding story' with NotebookLM

Akiyama uploaded the following documents lying dormant within the company:

• Representative interview articles

• Past presentation materials

• Initial press releases

Prompt: 'Extract words and episodes that represent the "reason for the company's existence" from these three documents, and analyze the commonalities and differences in expression. In particular, please visualize the connection between "individual passion" and "social issues."'

→ Rediscovered that the "soul of the founding" resides in words that previously seemed scattered. Realized the reintegration of PR and management.

④ Refining 'text that speaks about the mission' naturally with Claude

Prompt: 'I want to create a representative greeting that speaks about our mission, "Supporting Life." Please avoid excessive abstraction, weave in specific episodes and social changes, and make it a "conversational" text.'

→ Converted into "conversational language" that resonates with both external parties and employees. Can also be deployed for recruitment videos and IR documents.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt) | ChatGPT | Creating questions to delve into raison d'etre | "Design questions to delve into the meaning of the mission from 4 perspectives" | Genspark | Outputting diverse "raison d'etre cards" | "Express 'Supporting Life' in 10 ways for different targets" | NotebookLM | Extracting common stories from past internal materials | "Find the points of agreement in 'fundamental motives' across the founding period, initial PR, and representative's words" | Claude | Converting abstract mission into natural language | "Create a representative greeting that speaks about the mission using specific episodes + a conversational tone"

■ Column: A philosophy is not something you 'write,' but something you 'nurture'

Do you think, 'Now that we've decided on our mission, I have words I can use as PR'?

But in reality, a mission is just the starting point. It is something you nurture by rephrasing, testing, and retelling it over and over again.

Generative AI demonstrates its true value when used not to 'finalize words,' but to 'expand and deepen them.'

■ What should you do in these situations?

Q: Our mission is too abstract, and I'm having trouble explaining it
→ You can organize it by using Genspark to generate multiple rephrasing cards for different target audiences.

Q: There is no consistency between internal documents and external communications
→ You can identify points for improvement by using NotebookLM to highlight 'expression gaps.'

Q: Management says, 'Only those who understand the mission need to understand it'
→ Use ChatGPT to design questions that clarify the 'significance that resonates both internally and externally,' and use them as a starting point for internal dialogue.

■ Reader Workshop: 3 Steps to Approaching 'Corporate Purpose'

1. Design questions with ChatGPT to create opportunities for dialogue

2. Prepare 'ways of speaking for different targets' using Genspark

3. Reconstruct the 'founding vision' using NotebookLM

■ Conclusion: PR that nurtures words builds the 'backbone' of the company

A company's purpose is a promise to society and a promise to its employees.
The job of PR is to turn that into 'words that can be explained to anyone.'

Why don't you, together with your partner, Generative AI,
go and search once more for the words that build the 'core' of your company?

17 | The First Step in Startup PR

(Chapter 2 | Verbalizing Strategy - Building the 'Core' of PR)

■ Prologue: Stuck at 'Started Twitter'

Education startup EduGrow. In its second year of business, with 12 employees.
The product is a learning platform that uses personalized AI. They have been appearing at educational events, and attention within the industry is gradually growing.

In the midst of this, the person who became the company's first PR representative was former Customer Success member Aki Minami.
She was appointed by the CEO who said, 'Shall we start doing PR now?' but—

All she received were vague instructions.
Minami also didn't know “what she should be communicating in the first place,” and her posts stopped after three months.

■ Case: From the day she was told, “Do PR”

Minami was facing the “three major challenges of launching PR” common to startups:

• Being told to just post something without any strategy or KPIs

• The company itself doesn't understand “what PR is”

• Feeling like there is absolutely no content to post

However, looking at it from a different perspective, Minami had a “blank map for PR that she could design freely” from the start.
She opened ChatGPT and began to “design” through trial and error.

■ Solution: Drawing the “first map” with generative AI

① Inventory the “PR functions needed by the company right now” using ChatGPT

Prompt: “Please organize the ‘overall picture of PR functions’ for a company like the following: ・2nd year since founding ・12 employees ・Product is educational SaaS ・No marketing department ・Currently strengthening recruitment → Please identify the (1) PR (external communication), (2) internal communication, and (3) coordination with management that PR should handle.”

→ By breaking down “what needs to be done,” she could see “what to start with first.”

② Create a “PR statement (= her own flag)” with Claude

Prompt: “I want to verbalize ‘what kind of PR activities I will conduct as the company's first PR representative.’ The company's philosophy is ‘Make education freer.’ Based on this, please propose ‘the ideal form of PR I am aiming for’ in 200-300 characters.”

→ The vague idea of “what to aim for in PR” was refined intowords with a solid core.”

③ Organize “past presentation materials and sales materials” with NotebookLM to unearth content for posting

Minami uploaded the CEO's slides, sales materials, and PDFs of past event presentations.

Prompt: “Based on the uploaded materials, please extract ‘the story this company should tell’ in 5 patterns: (1) Founding story, (2) Customer testimonials, (3) View on education, (4) Product evolution, (5) CEO's beliefs.”

→ Far from “having no content,” she realized that the stories to tellalready existed.”

④ Build a “company introduction slide (for external use)” from scratch with Canva

Prompt: "Create a structure for a startup company introduction deck with the following sections: - Company Overview (1 slide) - Founding Story (1 slide) - Product Introduction (2 slides) - Case Studies/Achievements (1 slide) - Vision and Future (1 slide)"

→ From "I don't know what to show" to "I can talk about this!"The first presentation material for dissemination is complete.

■ Summary of AI Tools and Prompts Used

Tool | Usage Scenario | Prompt (Excerpt) | ChatGPT | Organizing PR roles/Visualizing tasks | "Organize the PR functions needed for a company in its second year into 3 categories" | Claude | Creating a "PR statement" that serves as your flag | "Express a 'PR policy' aligned with the company's philosophy in 200-300 characters" | NotebookLM | Turning materials into stories and discovering topics | "Organize the stories that should be told based on presentation and sales materials into 5 categories" | Canva | Creating company introduction materials for PR | "Propose a structure for a 'compelling company introduction slide' for a startup"

■ Column: The First PR Job is to "Plant a Flag"

When a dedicated PR person joins a startup, there is no single correct answer.

That is precisely whyplanting a flag that says 'this is the kind of PR I want to do'becomes the starting point for everything.

Generative AI can calmly organize the 'PR the company needs right now' and help you find the words to give your 'flag' persuasive power.

■ What to do in these situations?

Q: PR KPIs aren't set, and I'm struggling with what to aim for
→ Have ChatGPT generate priority KPIs based on the conditions: "2 years since founding, strengthening recruitment, small scale".

Q: The CEO says there are "no topics"...
→ Upload past presentation materials to NotebookLM for analysis. The "true feelings" expressed there are the topics.

Q: Where should I start with PR?
→ Ask Claude for the "first steps for a solo PR person in a startup" to draw a roadmap for action.

■ Reader Workshop: 3 Steps to Becoming a "Flag-Planting PR Professional"

1. Inventory the "PR functions needed by the company right now" using ChatGPT

2. Create the "PR language you aim for" using Claude

3. Shape your "first materials and story" with NotebookLM and Canva

■ Conclusion: Your "Flag" Will Guide All Your PR Efforts

The anxiety felt by Aki Minami is a path everyone who starts PR walks at least once.
However, once you verbalize "what you are aiming to communicate,"
meaning and policy will reside in every action.

Let's plant a flag. You can create that flag together with AI.

18 | How do you translate management's "impossible demands" into PR?

(Chapter 2 | Verbalizing Strategy - Creating the 'Core' of PR)

■ Introduction: 'How should this company be talked about?' is invisible

Logistics tech company MOVELOGIC. 10 years since founding, about 200 employees.
They are developing B2B SaaS in the logistics x AI domain and have entered a stage with an eye on an IPO.

Newly appointed as CMO (Chief Marketing Officer), Kurokawa said this at the board meeting.

Following that, the mission given to PR representative Yuka Mashita was as follows.

...Mashita, who has 5 years of PR experience, was also speechless.

Mashita took on the difficult challenge of
'translating into PR' the "thoughts" and "ideal images" of management that cannot be put into words, using real language, alongside AI.

■ Case: There is something to say. But it is not in a "sayable form"

Executives and management sometimes make "poetic" requests to PR.

• '"Logistics revolution" is weak. Use an expression more like "the backbone of society."'

• 'I want to give an impression that is "human," not "cutting-edge."'

• 'I want words to speak to society, not to customers.'

Mashita was troubled. How to proceed with the "translation" between abstract and reality, philosophy and practice.

■ Solution: Translating the "unspoken intentions" of management with AI

1. Using ChatGPT to "convert management statements into PR keywords"

Prompt: "The following are notes from management. Please convert this content into "keywords, hooks, and paraphrasing candidates that can be used for PR." Example) - Supporting social infrastructure - Designing the future of logistics - Philosophy over technology - Aiming for "transparent delivery"" -> Please present 10 paraphrasing suggestions each from a PR perspective, while inferring the intention and background of each.

-> Words that were too poetic begin to be concretized into "explainable words."

② Using Claude to "refine executive speech notes into a compelling narrative"

Prompt: "The following are notes from a CMO's internal presentation. Please refine the spoken content into a text that resonates with the reader. ・Create a 'transparent mechanism that supports the future of logistics' ・It should be 'logistics driven by human will' ・I want to keep asking 'what does it mean to deliver?' → The tone should be approachable, balancing abstraction and concrete examples."

→ Vague spoken words evolve into "compelling language" that can be used at the beginning of a PR article.

③ Using NotebookLM to "extract core concepts from past statements and documents"

Mashita uploads past internal presentation materials, interview transcripts, and business strategy documents from representatives and executives.

Prompt: "From the uploaded materials, extract expressions that represent 'why the company exists.' List common phrases, vocabulary, and impressive turns of phrase, and reconstruct them as the 'roots of our philosophy.'"

→ The gap between "unchanging thoughts" and "ever-changing expressions" is visualized, and
the company's 'linguistic axis' is redesigned.

④ Creating "vision sharing slides and vocabulary lists" with Canva and Notion

Generate "internal sharing slides" and a "list of words to use" that can be utilized in Slack or meetings.

Prompt: "Please propose the following to help share the company's vision and social significance with employees and new members in an easy-to-understand way: ① Slide structure plan (for Canva) ② List of 'words we want to use' (for Notion)."

→ A mechanism for creating 'words that can be spoken internally before being released externally' is complete.

■ Summary of AI tools and prompts used

Tool | Use Case | Prompt (Excerpt) | ChatGPT | 'Translating' poetic executive statements | 'Paraphrase the executive's abstract statements into 10 PR keywords each' | Claude | Refining statements into 'readable manuscripts' | 'Refine spoken statements into approachable reading material' | NotebookLM | Extracting 'roots of philosophy' across documents | 'Reconstruct "why we exist" from representative statement materials' | Canva / Notion | Creating internal sharing materials and 'common language' lists | 'Present a vision-sharing slide and a list of words to use as a structural plan'

■ Column: PR is a "translator" and an "interpreter"

An executive's words look too far into the future. Therefore, they need to be brought down to reality. PR is that "interpreter" and also an "editor."

AI is optimal as an auxiliary brain for translation work. It can move back and forth between "poetry" and "data" as many times as needed.

■ What should I do in these situations?

Q: The executive's statements are too abstract, and I don't know how to verbalize them.
→ Let's use ChatGPT and Claude to refine them with a set of "paraphrasing" and "natural tone."

Q: The company's language lacks consistency.
→ Let's re-excavate and organize the "common language of philosophy" using NotebookLM.

Q: Even within the company, people cannot articulate 'what this company is aiming for.'
→ Create 'speakable materials' and a 'word list' using Canva and Notion, and share them company-wide.

■ Reader Workshop: 3 Steps to "Translate the Executive's Words"

1. Use ChatGPT to rephrase abstract speech into "communicable keywords"

2. Use Claude to restructure it into a readable and speakable manuscript

3. Use NotebookLM to unearth the "roots of the philosophy" from speech materials

■ Conclusion: The Job of Converting Management's "Future Language" into Society's "Everyday Language"

Public relations is the "job of translating ideals into reality."

That is why, now that you have AI as a translation partner,
you can turn any "philosophy" or any "abstract term" into
the language of public relations.

19 | Internal Dialogue Techniques for When Public Relations Activities Are Not Understood

(Chapter 2 | Verbalizing Strategy - Creating the "Core" of Public Relations)

■ Introduction: Freezing at the remark, "So, what does PR actually do?"

Koki Solutions, a BtoB company supporting manufacturing DX.Founded 30 years ago, with about 500 employees.
Due to a sales-driven organizational structure, their order share within the industry has been growing year by year, but they have hardly engaged in branding until now.

Last year, the first full-time PR person wasYuki Miyake.Coming from an advertising agency background, while she was in charge of launching SNS and owned media, such voices were leaking within the company.

The anxiety and loneliness of not being understood.
Miyake decided to redesign PR to become a "speakable presence" within the company using AI.

■ Case: How to Respond to Internal "Indifference"?

PR work is hard to see. It cannot be measured by numbers alone.
However, if left as is, it will end up as a "department that no one knows what it does."

Miyake recognized the following three challenges.

• Not "connected" to other departments such as sales, development, and back office

• Perceived by the company as having "no visible results"

• Even they themselves have not clearly verbalized the significance of their activities

It wasn't “internal PR” that was needed, but “PR that makes the company understand PR.”

■ Solution: Using AI to refine the significance of PR through “words and structure”

1. Create a “dialogue scenario to visualize the connection between other departments and PR” using ChatGPT

Prompt: “In a B2B manufacturing company, please create a dialogue scenario for an internal meeting where PR can collaborate with other departments and gain their understanding. Target departments: Sales, Development, HR. Purpose: To convey how PR is useful from the ‘other party's perspective.’”

→ The “explanatory language” to be used in conversations with other departments was refined, and a dialogue blueprint for building relationships was completed.

2. Refine the “tone of voice” for PR activity reports shared internally using Claude

Prompt: “Please turn the following PR activity results into ‘easy-to-read and meaningful text’ for internal sharing. Performance examples: - SNS followers increased by 500 per month - 5 media placements (mainly industry trade papers) - 5 articles updated on owned media. → Please make the narrative convey value even to those who don't know PR.”

→ By providing a “context that conveys meaning” rather than just numbers, the activities were “visualized.”

3. Visualize the “relationship between past PR initiatives and internal reactions” using NotebookLM

Miyake uploaded internal newsletters, Slack posts, internal surveys, and sales accompaniment records.

Prompt: “From the uploaded internal documents, please extract trends on ‘how PR initiatives were perceived within the company.’ Classify them into: - Positive reactions - Negative reactions - No reaction. Please organize the elements that can be utilized in dialogue.”

→ Beyond just results, I was able to grasp “what was conveyed / what was not conveyed.”

4. Create “PR overview slides for internal sharing” using Canva

Prompt: “Please provide a structure for an internal presentation material that conveys ‘what PR is doing’ with the following composition: - Overall tasks handled by PR (1 slide) - What value it brings to the company (1 slide) - Recent achievements and episodes (1 slide) - Future goals and requests (1 slide).”

→ Materials that can properly explain the “existence of PR” were completed, and trust was fostered in internal meetings.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Used Prompt (Excerpt) | ChatGPT | Creating dialogue scenarios with other departments | “Create a ‘PR explanation meeting script’ for Sales, Development, and HR” | Claude | Refining the ‘tone of voice’ for internal shared documents | “Organize PR activities into a ‘meaningful form’ and make it an internal sharing text” | NotebookLM | Analyzing reaction trends to internal documents | “Classify trends in reactions to PR initiatives from internal newsletters and Slack” | Canva | Creating internal materials to explain the existence of PR | “Create a draft slide structure for internal use that answers ‘What does PR do?’”

■ Column: PR won't be understood unless you ‘advertise’ it internally as well

PR work is hard to see, especially within the company.
If you think ‘those who understand will understand,’ you will remain an ‘isolated department’ forever.

PR work also requires a proper ‘effort to convey’ it ourselves.
That means advertising ourselves within the company.

AI handles the groundwork "quietly and accurately."

■ What to do in these situations?

Q: I feel like PR is invisible within the company
→ Use Canva and Claude to prepare "slides and language that tell a story," and share them internally on a regular basis.

Q: I'm not building connections with other departments
→ Use ChatGPT to design "value communication from the other party's perspective."

Q: I can't quite grasp the internal reaction
→ Use NotebookLM to summarize past Slack messages and surveys, and have it analyze reaction trends.

■ Reader Workshop: "3 Steps to Communicate PR Internally"

1. Design a "dialogue script" for other departments using ChatGPT

2. Create internal documents that refine the "PR tone of voice" using Claude

3. Create a slide that conveys the "big picture of PR" in one page using Canva

■ Finally: The day will come when you are called "the PR person" from within the company

The materials Miyake created, and each explanation,
eventually led to messages on Slack saying, "Hey PR person, I'd like to consult with you about something."

Let's properly talk about your work.
Let's turn your presence into "meaningful words" together with AI.

20 | Can't get a PR budget... How to become someone who "can speak up"

(Chapter 2 | Verbalizing Strategy - Creating the "Core" of PR)

■ Introduction: Every time a presentation ends, I let out a sigh

The local startup "AgriLink" is a venture company gradually gaining attention for its agricultural support SaaS.
With 40 employees, there is only one person handling both marketing and PR. That person was the PR representative, Misa Nishina.

"The service is good. It has social significance, too.
But the money I can spend on PR... is almost nothing."

Conversations like this were standard within the company.

Nishina proposed PR initiatives to the board multiple times, but they were rejected every single time.
"I lack the power to convey the value of conveying."
—Nishina was keenly aware of that reality.

■ Case: To secure a budget, you must communicate the "necessity of PR" using "numbers and words"

Nishina's challenge was clear.

• Failing to show the management team how PR contributes to sales

• PR initiatives are discussed too much in terms of "feelings" and do not reach the language of decision-makers

• My proposals lack "numbers," "the future," and a "management perspective"

"If things continue like this, nothing will change."

Nishina used generative AI to start preparing to become someone who can speak effectively.

■ Solution: Build a "persuasive PR budget presentation" with generative AI

① Create a "PR initiative presentation draft from a management perspective" using ChatGPT

Prompt: "Please organize the following PR initiative proposal for a presentation to the management team. Objective: Strengthening recruitment PR / Increasing awareness / Building media relations. Target: CEO and CFO who decide the PR budget. Conditions: Make the structure logical and easy to decide on in a short time. Elements: ① Current issues ② Initiatives ③ Required budget ④ Expected effects ⑤ Risk hedging"

→ Initiatives spoken in "management language" changed into words that get heard.

② Refine "numerical backing + narrative tailored to decision-maker interests" with Claude

Prompt: "Please convert the following materials into an 'explanatory text that moves decision-makers.' Points: ・KPIs (e.g., number of articles published, SNS CVR) ・Benchmarks (competitors) ・Risk explanations that avoid misleading ・Presentation of ROI hypotheses"

→ Able to speak logically based on numbers, not feelings.

③ Analyze "past initiatives and internal decision-making trends" with NotebookLM

Nishina uploaded past budget meeting materials, approval memos, and executive comments.

Prompt: "Extract the characteristics of presentations that easily get budget approval from past decision records and meeting materials. Organize the trends, vocabulary used, structure, and perspectives emphasized, and derive a success pattern"

→ The company's "language patterns that lead to decisions" became clear.

④ Visualize "budget allocation plans and KPI setting tables by initiative" with Copilot

Prompt: "Create a budget allocation plan, KPIs, and outcome prediction graphs for the following PR initiatives. Format: In a table format that can be shown in Excel/PowerPoint"

→ Because the "presentable format" was ready,the approval rate in meetings increased dramatically.

■ Summary of AI Tools and Prompts Used

Tool | Use Case | Prompt (Excerpt)
ChatGPT | Creating presentations that resonate with management | "Create a draft explaining PR budget measures for the CEO and CFO"
Claude | Polishing into "persuasive text" that speaks with numbers | "Refine the explanation of PR measures, including ROI hypotheses and benchmarks"
NotebookLM | Discovering "approvable explanations" from past approval patterns | "Extract commonalities from presentations that secured budget and organize success trends"
Copilot | "Visualizing" KPIs and budget proposals | "Propose budget allocation plans, KPI settings, and outcome forecasts for each measure in a table format"

■ Column: Persuasive people speak in the "other party's language"

If you cannot talk about "what PR becomes for management" rather than what PR does, you will not get a budget.
You must be able to articulate this.

Generative AI acts as a bridge to "translate"from your own language to the language of management.Let's have words that "move" people, not just "convey" information.

■ What to do in these situations?

Q: I can't speak well in budget meetings...
→ Use ChatGPT to create a "speaking template for decision-makers" and practice.

Q: I don't really understand why it wasn't approved in the past
→ Use NotebookLM to analyze past meeting materials and extract "patterns for successful presentations."

Q: I'm not good with numbers and don't really understand ROI
→ Use Claude to organize benchmarks and hypothesis logic, and use Copilot to create "visual materials."

■ Reader Workshop: "3 Steps to Becoming Someone Who Can Talk About Budget"

1. Design a "structure that reaches decision-makers" with ChatGPT

2. Refine into an explanation that "speaks with numbers and logic" using Claude

3. Summarize "visual" materials and KPIs in a table with Copilot

■ Finally: Once you can say it, it will get approved

Nishina realized something.
When she became able to say, "We need X yen for Y," instead of "We need PR,"
PR was finally recognized as a "weapon for the organization."

By becoming someone who can "speak up," PR becomes a force that builds the company's future.

21 | I want to make the President's message resonate!

(Chapter 3 | Relationship Building and Empathy - Dialogue Skills with Internal and External Stakeholders)

■ Introduction: The "President's words" do not resonate with employees or the media

IT infrastructure company 'Cloud Gear'. It has 200 employees. As it continues to grow year by year, organizational fragmentation and the normalization of remote work have progressed, leading to increasing comments both inside and outside the company that 'the face of the company is becoming harder to see'.

Amidst this, the public relations representative, Madoka Sakata, faces a certain event.

For the 10th anniversary of the company's founding, she received a request from the president to 'send a message to the employees'.
She finished the draft and felt she had packed in the president's thoughts, but—

• Internal survey: 'It didn't leave much of an impression'

• SNS: 'Sounds like empty platitudes', 'Who is this aimed at?'

• Media: 'It might be difficult to publish... it's a bit abstract'

Sakata decided to use generative AI to reconstruct a 'president's message that generates empathy'.

■ Case: Not just 'polishing' the president's words, but making them 'worth listening to'

Sakata organized the following issues.

• The president's message is abstract and not reaching anyone

• There is no distinction in tone between internal and external use

• The original 'personality' and 'passion' are not coming through in the text

It was not a problem of expression, but a problem of design.

■ Solution: Redesigning words to connect generative AI with the 'president's reality'

① Create a 'conversational interview draft that exudes the president's character' using ChatGPT

First, change the format. Have the president speak in an interview format and transcribe it directly into text.

Prompt: 'Based on the following questions, please structure a message for employees in the format of a president's interview. Example questions: - What did you value most when you started the company? - What 'employee action' has made you happiest recently? - What kind of company do you want to build in the future? -> Please structure it in a tone that conveys his 'character'.

-> The 'passion of the speaker' seeped through the text, and voices within the company rose saying, 'I'm glad I read it'.

② Use Claude to convert 'abstract words' to 'concrete words' and refine them into 'visible words'

Prompt: "Please rewrite the following draft of the president's message, avoiding overly abstract expressions and using specific episodes or action-based language. Objective: To make the words 'conveyable and relatable' to employees and readers."

→ The 'thoughts' and 'philosophy' became words that connect with employee actions.

③ Organize 'past message documents and trends in internal reactions' using NotebookLM

Analyze past New Year's greetings, company-wide speeches, PR articles, and social media posts collectively.

Prompt: "From past communications by the president (articles, social media, greetings), please extract trends in words, expressions, and structures that received a response inside and outside the company."

→ A 'pattern that resonates,' such as 'mentioning employees by name' and 'describing scenes rather than using numbers,' became visible.

④ Create 'visual word posters' using Canva

Prompt: "From the following president's message, please create a poster-style visual for internal posting and distribution. - Summarize into 3 points - Include the president's photo + famous quote + handwritten-style elements"

→ The visual impact was strengthened, and the share rate on Slack increased by more than 5 times.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt) | ChatGPT | Creating messages in a president interview format | "Create a dialogue structure that utilizes the 'president's unique way of speaking'" | Claude | Converting abstract expressions to concrete ones | "Convert the philosophy into concrete words that lead to 'employee actions'" | NotebookLM | Analyzing communication patterns that resonated | "Extract 'words and structures that resonated' from past communications" | Canva | Designing 'word posters' for internal use | "Create a poster with a visual summarized in 3 points + the president's photo"

■ Column: The way 'thoughts' are received changes depending on the 'order of delivery'

If you talk about 'philosophy' or 'thoughts' first, people think, 'Here we go again.'
However, if you speak in the order of event that resonated -> background -> thoughts, the way it is received will be completely different.

AI is an editor that designs where to start speaking to make it understood together with you.

■ What to do in these situations?

Q: The president only says abstract things...
→ Let's design a 'question scenario that resonates with employees' using ChatGPT and conduct it in an interview format.

Q: The message feels too polished and lacks emotion
→ Let's convert 'moments that moved the heart' into concrete words using Claude.

Q: It's not read even after distributing it internally...
→ Let's turn it into visual material using Canva and link it with Slack or internal newsletters.

■ Reader Workshop: 3 Steps to a 'President's Message that Resonates'

1. Draw out the president's words in a 'question format' using ChatGPT

2. Translating abstract words into "emotional movements" with Claude

3. Visualizing with Canva to create a "memorable" message

■ Finally: Let's deliver the CEO's "thoughts" with your PR power

Sakata says.

That is the power of empathy you create.
Let's turn the CEO's thoughts intowords that reach the hearts of employees.

22 | Employee interviews: How much should be disclosed?

(Chapter 3 | Relationship Building and Empathy - Dialogue Skills with Internal and External Stakeholders)

■ Introduction: "They spoke so candidly, but... is it okay to publish this?"

IT venture "Linkware."
120 employees. A PR team of two.

Shota Sakai,in his second year of PR,was working on his firstemployee interview article.

He interviewedEngineer Yoshida,a third-year employee.
Passion for work, stories of past failures, frank thoughts about the company—
The interview was lively, and many real, moving episodes emerged.

However, as Sakai finished writing the draft, these anxieties swirled in his head.

• "Is it okay to publish this failure story...?"

• "Is it too honest and might it give a negative impression of the company?"

• "How much should I show to people inside and outside the company...?"

The article is "true-to-life and real"but at the same time,"raw and a bit scary."

Sakai stopped moving, his hands frozen in front of his computer with the draft open.

■ Case: Objectively viewing the "publish/don't publish line" with AI

Troubled, Sakai first entered the following prompt intoClaude.

Prompt: "Please review the following employee interview draft from the perspective of whether it is appropriate for public release. - Balance of positive and negative expressions - Consideration for internal and external stakeholders - Impact on corporate image - Personal information risks - Legal and ethical considerations. Based on this, please specifically point out any risky areas and suggest improvements."

As a result, the following improvement points were output.

• Risk 1: Excessively negative failure stories → "Convert to expressions that conclude with positive learning"

• Risk 2: Phrasing that appears to criticize the company → "Add supplementary remarks that emphasize the company's stance on improvement"

• Risk 3: Sensitive expressions using real names → "Mask real names or obtain explicit consent from the individual"

Furthermore, I used the following prompt in ChatGPT to adjust it to an "engineer recruitment article tone."

Prompt: "Please rewrite the following interview article for 'engineer recruitment candidates,' keeping it realistic while maintaining a positive and future-oriented tone."

Finally, I used Notion AI to organize headings and subheadings, enhancing readability.

■ Solution: An article that balances 'honesty' and 'peace of mind'

The final draft turned out as follows.

• Keep the authentic personal experiences as they are

• However, always conclude after reading with "positive learning" and a "growth story"

• For expressions that might place a burden on an individual, always confirm with the person in advance

After publication, it was well-received internally, with comments like "It's realistic and great" and "It conveys the company's atmosphere!"
Sakai herself reflects, "By having AI provide an objective review with a 'cool-headed perspective,' I gained peace of mind as a PR professional."

■ Summary of AI Tools and Prompts Used

Tool Usage Scenario Used Prompt (Excerpt) Claude Risk check/objective review "Review the employee interview draft from the perspective of 'is it okay for public release' and point out risky areas" ChatGPT Tone adjustment/rewrite "Please rewrite for engineer recruitment candidates in a realistic and positive tone" Notion AI Heading/subheading organization "Please add appropriate subheadings to this draft and organize it for readability"

■ Column: "What lies between 'candor' and 'PR responsibility'"

The appeal of employee interviews lies in their "authentic voices."
However, PR professionals have a "responsibility to deliver them safely from the reader's perspective."

Honesty is a weapon.
But being defenseless is a risk.

AI acts as a 'cool-headed filter' that bridges that gap.

■ What to do in these situations?

Q: An employee made a 'bombshell statement'...
→ Use a 'risk phrase extraction prompt' in Claude to automatically identify areas needing attention.

Q: I want to summarize it positively, but I don't want to overstate it.
→ Input a prompt into ChatGPT to 'summarize in a positive tone without changing the facts'.

Q: The tone is inconsistent across multiple employee interview articles...
→ List past articles in Notion and ask Claude to 'edit for tone consistency'.

■ Reader Workshop: Let's try 'Risk Checking Your Company's Interview Articles'

1. Upload past employee interview drafts to NotebookLM or Claude

2. Perform an objective review using an 'external disclosure risk check' prompt

3. Create a tone-adjusted version in ChatGPT as needed

4. Refine subheadings and structure in Notion to create the final draft

■ Conclusion

Employee interviews are the best content for conveying the 'vibe of the company.'
But at the same time, it is a genre that demands responsibility and consideration from PR professionals.

AI is a 'calm partner' for fulfilling that responsibility.
The perfect balance between 'authenticity' and 'peace of mind'—
please continue to explore this with AI.

23 | How can you convert management's words into a 'usable' format?

(Chapter 3 | Relationship Building and Empathy — Dialogue Skills with Internal and External Stakeholders)

■ Introduction: 'The problem where the president's words cannot be used as-is when drafted'

Central Products, a mid-sized manufacturing company.
450 employees, 50 years in business.

PR representative Shohei Kobayashi was worried every time.

The mission this time was the president's new fiscal year policy message.

• 'Please summarize it into a script for a video message to employees'

• 'Make use of the president's own words, and make it feel like our company culture'

But—.

When you actually transcribe a recorded interview with the CEO...

• Circumlocutory expressions

• A barrage of business jargon

• Disconnected context

• Ambiguous subjects

• The "So, what are you trying to say?" state

Kobayashi held his head in his hands at his desk.

■ Case: "Polishing the CEO's 'raw gems of speech' with AI"

Kobayashi first uploaded all of the CEO's past message drafts and meeting minutes to NotebookLM.

Then, he entered the following prompt.

Prompt: "From the following documents, create a list of 'expressions and catchphrases this CEO often uses.' Furthermore, create examples of how to convert these into 'words that resonate with employees.'"

Next, he pasted the transcription of the recorded data into Claude and requested a rewrite with the following prompt.

Prompt: "Please restructure the following executive interview draft for a video message to employees. - Positive and forward-looking tone - Length readable within 3 minutes - Easy-to-understand word choice - Expressions that field employees can empathize with"

Furthermore, he had ChatGPT output multiple patterns for 'opening catchphrase ideas' and 'closing remark phrases.' Finally, he used
Notion AI to organize the paragraphs and polish the text for narration.

■ Solution: 'Words that couldn't be used as-is' turned into 'messages that get across'

The final draft hit the following points.

• While keeping the CEO's catchphrases and mannerisms, they were converted to be easy to understand from an employee's perspective

• The flow became smooth with a three-part structure: introduction -> main message -> conclusion

• "Ready-to-use phrases for the field" were placed throughout, and they even started being quoted on the internal social media.

Kobayashi realized, "By incorporating AI, my mindset shifted from 'fixing' to 'leveraging'.

■ Summary of AI Tools Used and Prompts

Tool | Usage Scenario | Used Prompt (Excerpt)
NotebookLM | Analyzing past statement patterns of the president | "From this set of documents, list the 'expressions the president prefers' and provide easy-to-understand conversion examples."
Claude | Long-form rewriting and creating easy-to-understand scripts | "Reconstruct this into a message for employees under 3 minutes, using a positive and empathetic tone."
ChatGPT | Generating catchphrases and closing phrases | "Please provide 5 options each for an opening catchphrase and a closing remark that fit this message."
Notion AI | Editing as a narration script | "Please re-edit this text with a 'short-sentence rhythm that is easy to read aloud'."

■ Column: "'President-likeness' shouldn't be cut, but 'refined'"

The job of public relations is not just to "convey the words of management."
It is to "convert them into a form that gets across" and deliver them.

Words in their raw state are rough.
By passing them through an AI, which acts as a "word polisher,"
they become "words that reach the field" for the first time.

"Refining is not a betrayal."
It is, in fact, "an act of public relations protecting the executive's vision."

■ What to do in these situations?

Q: The president says, "Just release the draft as it is..."
→ Try using a prompt in Claude to "keep the draft as is, but refine the structure to hit home with the audience."

Q: It's too long and employees won't read it.
→ Instruct ChatGPT to "summarize this to a length employees can read in under 2 minutes, and narrow it down to 3 key points."

Q: There are too many 'management terms' and the field can't keep up.
→ Use NotebookLM to create a "list of easy-to-understand terminology conversions" for field employees.

■ Reader Workshop: "Workshop to convert management's words into a 'usable form'"

1. List past management messages using NotebookLM

2. Convert tone for employees using Claude

3. Create "opening catchphrases" and "closing phrases" using ChatGPT

4. Organize into an easy-to-read script using Notion

■ In Conclusion

In the words of management,
"the future of the company" resides.

Just releasing them as they are won't get them across.
But by having public relations refine them,
the "warmth of the words" will surely be conveyed.

24 | Recruitment PR: Where should I start?

(Chapter 3 | Relationship Building and Empathy - Dialogue Skills with Internal and External Stakeholders)

■ Introduction: "How involved should HR and PR be?"

Nature Foods, a food startup.
60 employees, 7th year since founding.

Kenichi Maehara, Kenichi Maehara who has been in the PR department for six months after transferring from marketing, was assigned his first "recruitment PR" task.

However, he immediately hit a wall.

• "To begin with, where should I start with recruitment PR?"

• "HR is already creating job postings and scout emails, so where should PR get involved?"

• "What is the difference between recruitment articles and regular PR articles?"

Within the company, there were voices saying "I want to improve the recruitment site" or "Let's increase employee interviews," but
no concrete policy was in sight.

■ Case: "Connecting information from the candidate's perspective"

The first thing Maehara did was research using Perplexity.

Prompt: "Please tell me about effective recruitment PR initiatives at 5 startup companies. I would especially like to know about examples of utilizing employee interviews and recruitment blogs."

Next, he uploaded all of the company's past recruitment-related content (job postings, internal blogs, past interview articles) to NotebookLM.

He then entered the following prompt.

Prompt: "From the following company recruitment content, please list the information that is lacking from a candidate's perspective. Also, please suggest necessary content themes for each persona."

As a result, the following issues were highlighted.

• "There is little content that conveys the company culture"

• "Employee interviews are too 'sugar-coated'"

• "Concrete examples of the growth environment and work style are lacking"

Next, I used the following prompt in ChatGPT to generate a "recruitment blog outline."

Prompt: "Please create 5 outline patterns for employee interview articles that would make candidates think, 'I want to work at this company.' The persona is a candidate in their 20s aspiring to be an engineer, who is considering a job change with a focus on job satisfaction."

Furthermore, I created rough drafts for recruitment page visuals in Canva, and created a "Recruitment PR Prompt Template Collection" in
Genspark to standardize the subsequent content production flow.

■ Solution: Able to design the entire process from a "candidate's perspective"

Ultimately, what Maehara discovered was:

• Recruitment strategy (target setting/persona) is led by HR

• Language design and information structure are led by PR

• Interview content and tone are co-designed by both parties

As a result,

• A consistent recruitment page from the candidate's perspective was completed

• Employee interviews achieved a balance of "authentic voices + company messaging"

• The division of roles between PR and HR became clear

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Used Prompt (Excerpt) | Perplexity | Competitor case research | "Please tell me about recruitment PR cases for startups." | NotebookLM | Analysis of in-house recruitment content | "Please list perspectives that are missing from a candidate's point of view." | ChatGPT | Recruitment blog outline creation | "5 outline patterns for interview articles that resonate with engineer candidates." | Canva | Recruitment page visual creation | "Create banner image ideas for an engineer recruitment page." | Genspark | Template creation | "Please create a collection of recruitment PR prompt templates."

■ Column: "Recruitment PR is the work of conveying a 'company's personality'"

People are not moved by job descriptions and conditions alone.

"What kind of people work here, and with what passion?"
"What kind of future awaits at this company?"

Conveying that atmosphere is the essence of recruitment PR.

AI organizes information, turns it into a story, and optimizes it for personas.
However, you are the one who decides, "Who do I want to envision what kind of future?"

■ What should I do in these situations?

Q: HR told me, 'I'll leave everything to PR'...
→ Use Genspark to create a 'PR and HR Role Division Template' and start by clearly defining the scope of work.

Q: Employee interviews are inconsistent every time...
→ Create an 'Interview Question List' in Notion and ask Claude to edit for consistent tone.

Q: Our recruitment blog isn't being read at all.
→ Use ChatGPT to generate 'high-engagement headline ideas and CTA (application flow) proposals' and adjust them.

■ Reader Workshop: Let's create your 'Company Recruitment PR Map'

1. Upload your company's recruitment page and past articles to NotebookLM

2. Use ChatGPT to generate 'content ideas needed for each persona'

3. Use Genspark to create a 'Recruitment PR Prompt Set' for your company

4. Create rough visual drafts for the recruitment page in Canva and share them with the team

■ Conclusion

Recruitment PR is a
love letter to the company's future colleagues.

Who are you sending it to?
What words will you use to deliver it?
Deciding that is not up to AI, but up to you.

25 | How to restructure a recruitment site from a candidate's perspective

(Chapter 3 | Building Relationships and Empathy - Dialogue Skills with Internal and External Stakeholders)

■ Introduction: 'Could the reason you aren't getting good candidates be your site?'

Metroware, a software company that spun off from a long-established manufacturer. 120 employees.
It is growing rapidly with its unique business support SaaS, but securing talent for both development and sales was a challenge.

PR representative Takuma Nomura was one day told this by a recruitment agent.

Indeed—

• There are no voices of employees or glimpses of office life

• It focuses on 'explaining systems,' failing to convey daily life or culture

• It was hard to read on smartphones, and the navigation was like a maze.

Could the site, which is the "gateway to recruitment," actually be something that makes people not want to enter?
Nomura decided to use the power of generative AI to fundamentally rethink the site.

■ Case: Changing the "face" of the site gets applications moving

Nomura first used ChatGPT and Perplexity to collect examples of "excellent recruitment sites from other companies."
Next, he uploaded his company's site structure, past applicant feedback, and entry data to NotebookLM.

Prompt (Perplexity): "Please tell me the structure, features, and success factors of companies with 'high application rates' on their recruitment sites for mid-sized IT companies." Prompt (NotebookLM): "Based on this site structure and past applicant feedback, please analyze where the bottlenecks are. Please evaluate the candidate's behavioral flow and emotional flow separately."

→ The issues were clear.

• Lack of smartphone-first design

• Content is not separated for engineers and sales staff

• Too many page transitions, leading to a high bounce rate

■ Solution: User-centered design + AI-edited reconstruction process

Nomura proceeded with the reconstruction using the following process:

1. Created a draft of the new structure using ChatGPT
   → Clearly branched the flow by persona (development applicants / sales applicants)

2. Tuned the manuscript for emotional flow using Claude
   → Structured it starting with episodes that make people feel, "I want to work with these people"

3. Organized and edited employee interviews and FAQs using Notion AI
   → Visualized "buried voices" within the company in an easy-to-read format

4. Created smartphone-optimized visuals and section banners using Canva
   → Designed it so that the atmosphere is conveyed even when "skimming"

As a result, the entry rate increased to 2.4 times the previous month within one month of the site's launch.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Used Prompt (Excerpt)
ChatGPT | Designing a new recruitment site structure | "Please create a structure proposal divided by candidate type (development/sales). Use a flow that serves as a starting point for empathy."
Perplexity | Case study research of other companies' recruitment sites | "Please structure and explain the features of excellent recruitment sites for SaaS companies."
NotebookLM | In-house site analysis and issue extraction | "Based on the site structure and feedback, please identify the reasons for candidate drop-off."
Claude | Empathy tuning of the manuscript | "Please restructure the following manuscript as a 'story of the people who work here.' Order it in a way that moves the reader's emotions."
Notion AI | Editing and organizing interviews and FAQs | "Please edit this employee interview to make it easy to read, and also create a Q&A format that can be repurposed for FAQs."
Canva | Creating smartphone-optimized visuals | "Please create smartphone-first banners and section images for the recruitment site."

■ Column: A recruitment site is your company's "customer service counter"

Who, with what expression, and what do they say—.
A recruitment site is an "unmanned service counter" that determines a company's first impression.

From structural design and word order to colors and white space.
Everything becomes a form of 'hospitality' for candidates.

Generative AI acts as a navigator that simulates the candidate's feelings,
accurately conveying the 'vibe' of your company.

■ What to do in these situations?

Q: People leave when they see our recruitment site on their smartphones because it's hard to read.
→ Use Canva to generate a 'mobile-optimized layout' and revamp your visual design.

Q: Employee interviews are scattered throughout the company and are difficult to use.
→ Use Notion AI to manage and summarize them in one place, then reclassify them by persona.

Q: I feel like it's unclear who the site is for, and the message isn't getting across.
→ Use ChatGPT to organize the user flow for each candidate type and design branching opening catchphrases.

■ Reader Workshop: Checklist for a 'Recruitment Site That Makes People Want to Apply'

1. Have ChatGPT break down and analyze your company's recruitment site structure.

2. Use NotebookLM to interpret past candidate feedback.

3. Rewrite copy to 'follow the emotional journey' using Claude.

4. Optimize banners and sections for each candidate type for mobile using Canva.

■ Conclusion

A recruitment site is a 'place for dialogue' that attracts people even when you aren't actively recruiting.
But to achieve that, you need the 'candidate's perspective' in every aspect: structure, words, and appearance.

You don't have to carry that perspective all by yourself.
Together with generative AI, let's simulate the candidate's journey and design a 'company people want to visit again'.

26 | Information leaks by employees... How to design guidelines

(Chapter 3 | Building Relationships and Empathy - Dialogue Skills with Internal and External Stakeholders)

■ Introduction: Unintended 'tweets' can become a major risk

The cosmetic D2C brand 'Rafale'.
A startup that gained popularity among Gen Z on social media, where employees also actively used X (formerly Twitter) and Instagram.

One day, PR representativeRena Shirakawawas called in by the president.

The person involved had no malicious intent.
A partnership with a partner company was almost finalized, and it seems they wrote it with the nuance of 'I'm looking forward to it!'

However, the single fact that it was 'before the official announcement' made it a leak of confidential information.

'It is no longer an era where only the PR department manages information.'
Shirakawa realized this keenly and began moving to codify the 'boundaries' that should be shared across the entire company.

■ Case: Designing 'Judgment Criteria' instead of 'Prohibitions'

First, Shirakawa usedChatGPTto identify the design items for the SNS guidelines.

Prompt: 'Please provide a list of items necessary to design SNS guidelines for a startup company. I want to provide 'judgment criteria' rather than 'prohibited items'.'

Next, Shirakawa uploadedNotebookLMwith examples of other companies' guidelines, past internal posts, and crisis response logs.

Prompt: 'From the following past cases, extract the 'causes that made information dissemination a risk' and 'perspectives to prevent it in advance.' Please propose '3 steps of judgment' that employees can use when they are unsure.'

→ Shirakawa decided on a policy of designing around '3 questions for when you are unsure' rather than a 'prohibition-based' approach.

Furthermore, Shirakawa usedClaudeto refine the tone of the text, finishing the draft while valuing the premise that 'we trust the employees' actions.'

Finally, Shirakawa usedCanvato create a 'visualized SNS judgment chart that can be understood in one page' and deployed it within the company.

■ Solution: 'Trust-based rules' protect organizational culture

Through this response, the quality of information management improved without dampening the employees' enthusiasm for dissemination.

• Employees voiced, 'It's reassuring to know what is okay up to this point.'

• It was well-received by managers as well, who said, 'It has become easier to explain.'

• The operation of the guidelines became a 'common language' rather than a 'sermon.'

Shirakawa usedGensparkto create a template and organized it so that it could be shared with other departments as an 'Information Dissemination Judgment Guide.'

■ Summary of AI tools used and prompts

Tool Usage Scene Used Prompt (Excerpt) ChatGPT Identifying SNS guideline design items 'Please create a structure to convey 'judgment criteria' rather than prohibitions.' NotebookLM Analysis of past cases and post logs 'Structure cases where information dissemination became a risk and organize 'perspectives to avoid hesitation.'' Claude Tone adjustment and preparation of drafts emphasizing trust 'Please make the guidelines gentle, using a tone that conveys trust in employees.' Canva Chart/Visualization 'Please make the SNS judgment chart a document that can be seen in one page so that employees can judge immediately.' Genspark Template structure/Organizational deployment 'Please create an 'SNS judgment template' that can be used across departments.'

■ Column: How to deliver rules with 'empathy' rather than 'caution'?

Guidelines exist not to 'control,' but to 'build trust and move forward together.'
Actions that employees take without malicious intent cannot be prevented by 'prohibitions' alone.

Generative AI can organize these gray zones based on case studies,
and tune the language so that 'hesitation' can be turned into 'decisions.'

The role of PR is not to stop people through fear, but to guide them through understanding.
That perspective leads to PR power that protects culture and fosters empathy.

■ What should I do in these situations?

Q: Employees don't read the guidelines because they are full of prohibitions.
→ Use Claude to convert them into a 'trust-based tone' and use Canva to format them into an easy-to-read design.

Q: Examples of SNS posts are scattered throughout the company and aren't being utilized.
→ Centralize them with NotebookLM and use ChatGPT to reconstruct them into 'common NG patterns' and 'safe examples.'

Q: I often get asked, 'Is it okay to say this?'
→ Use Genspark to develop a '3-step decision checklist' so that anyone can make decisions without hesitation.

■ Reader Workshop: Designing 'Decision Criteria for Employees to Post Without Hesitation'

1. Generate a 'list of decision criteria' and 'draft explanations' using ChatGPT.

2. Visualize the company's posting history and examples using NotebookLM, and organize the background behind the decisions.

3. Refine the text with Claude and convert it into a tone that is 'convincing and readable.'

4. Create a handout as an 'SNS Decision Chart' using Canva.

■ Conclusion

SNS guidelines are not meant to suppress communication.
Rather, they are like a common map to 'protect the space where you can speak in your own words.'

To draw that map, the power of PR is necessary.
Together with generative AI, let's cultivate 'language that enables decision-making' within the organization.

27 | How do you communicate cultural integration after a corporate merger?

(Chapter 3 | Relationship Building and Empathy — Dialogue Skills with Internal and External Stakeholders)

■ Introduction: Two corporate cultures that 'won't mix'

The long-established logistics company 'Asahi Line' and the DX support startup 'Spinalogics' completed a management integration following a business alliance.

The new company, 'Aspina Trans,' has 300 employees.
The trust between the top management is strong, and there is no gap in their future vision.

However, on the ground...

• Asahi Line side: "I feel like I have to be careful and can't speak my true feelings."

• Spinalogics side: "I don't know who to ask about the new system."

PR Director Takashi Takagi was struggling with how to communicate and resolve the 'temperature difference' in each side's values and atmosphere.

■ Case: Verbalizing the 'Story of Cultural Fusion'

To convey cultural differences, Takagi decided to start with a 'story' rather than 'regulations'.

First, he used NotebookLM to read both companies' internal newsletters, representative messages, and vision documents to extract the core of their organizational cultures.

Prompt: 'Compare the following texts and extract the characteristics of each company's culture, values, and language usage. Also, identify common ground that could serve as a "common language after integration."'

Next, he used ChatGPT to create a draft of the fusion story.

Prompt: 'Create an internal story that depicts the background of the integration of these two companies and the cultural gap felt on the ground, conveying the attitude that "we are now becoming one."'

Based on the story obtained, he used Claude to refine the text into a natural and soft tone, and Canva to visualize a 'cultural timeline' and 'employee voices'.

Furthermore, he used Genspark to generate a template for a 'Cultural Fusion Guide' to prepare for future deployment across departments.

■ Solution: Turning differences into a 'story' instead of a 'problem'

By verbalizing the fact that 'there are differences' instead of forcing the cultural gap to be closed, conversations between employees were sparked.

• The 'differences in what is taken for granted' between us and them became a topic of conversation in each department.

• The 'story of fusion' began to be told in representative messages as well.

• Respect was born by learning the 'reasons for each other's strengths'.

Takagi realized that what PR should fulfill is not 'explanation' but 'translation', and that generative AI is the ultimate tool for supporting that translation.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt) | NotebookLM | Analysis of both companies' materials and extraction of common values | 'Compare the vision and system documents of the two companies and extract cultural differences and common ground.' | ChatGPT | Drafting the fusion story | 'Create an internal story themed around the cultural fusion of the two companies.' | Claude | Tone adjustment and shaping into a style that evokes empathy | 'Finish it with a soft, conversational, and approachable tone.' | Canva | Production of visual materials (cultural timeline, employee voices) | 'Create internal materials that clearly organize the cultural timeline and employee voices of the two companies.' | Genspark | Generation of a template for the cultural fusion guide | 'Output a template for a "Cultural Fusion Guide" that can be deployed by department.'

■ Column: Integration is not 'reconciliation' but 'co-creation'

Internal communications after M&A or integration tend to end up as mere explanations of systems or formal greetings.

However, what employees are looking for is the narrative of 'why this integration was necessary,' and the articulation of a vision regarding 'who we are becoming.'

By using generative AI, you can reconnect the words of management with the voices of employees and reconstruct them as a story.
reconstruct them as a story.

Cultural fusion is not about eliminating differences.
It is about understanding, choosing, combining, and co-creating new value from those differences.

■ What to do in these situations?

Q: The cultures are fragmented and there is no common language
→ Use NotebookLM to compare documents and ChatGPT to extract 'keywords of shared values'.

Q: Employees are confused by the differences, but the atmosphere makes it hard to point out
→ Use Claude to convert it into a 'narrative format' and turn it into a story that delivers the message gently.

Q: Employees are anxious about what will happen next
→ Use Canva to visualize a 'cultural timeline' or 'future story' and depict a vision of walking together.

■ Reader Workshop: Creating a 'Cultural Fusion Guide'

1. Upload internal documents and vision statements to NotebookLM for analysis

2. Generate a draft of a 'narrative-style internal message' using ChatGPT

3. Refine it into a tone that resonates with employees using Claude

4. Design deployment templates with Genspark and visualize them with Canva

■ Conclusion

Public relations is the 'interpreter' when an organization is reborn.

Accurately grasp the differences in culture, systems, and language, determine where to build bridges, and deliver words that make each employee think, 'I want to live in this organization.'
deliver words that make each employee think, 'I want to live in this organization.'

Generative AI is a powerful ally that supports that translation and design.
Use the power of narrative to paint hope amidst change.

28 | Anniversary Projects: Communication Techniques That Don't End at 'Just a Celebration'

(Chapter 3 | Relationship Building and Empathy - Dialogue Skills with Internal and External Stakeholders)

■ Introduction: A milestone, yet no excitement?

Kobatech, a long-established manufacturer of machine parts, is a small-to-medium enterprise celebrating its 50th anniversary.
With 120 employees and a focus on B2B, their usual PR activities are minimal.

The PR representative, Go Ishida, was instructed by the president to "do something for PR since it's the 50th anniversary."

However, the internal reaction was cold.

• “A celebration? What’s the point of doing that?”

• “The shop floor is busy; we don’t have time for that.”

• “Who’s going to read an anniversary commemorative magazine?”

Ishida thought, "Perhaps this isn't just a 'celebration,' but a 'chance to reconstruct our story.'"

■ Case: Structuring the anniversary to communicate 'meaning'

Ishida first had ChatGPT read the company history, the president's message, and product materials to find a starting point for the story.

Prompt: "Based on the following content, propose a structure that conveys the company's history not as a 'celebration,' but as 'the succession and evolution of our philosophy.' Design it with a tone that communicates the 'future image of the company' to both internal and external audiences."

Next, he used Claude to rewrite episodes from the founding period and the words of the previous president into a dramatic format.

Furthermore, he used NotebookLM to analyze past recruitment brochures and sales materials, extracting keywords that connect the 'past' with the 'future.'

Canva was used to visualize the timeline and the evolution of the vision, and Genspark was used to design a 'template for the anniversary special page.'

Anniversary content was completed, serving not as a 'celebration' but as a 'milestone of evolution.'

■ Solution: Communicating as a 'milestone in the story' rather than a 'commemorative day'

As a result, the 50th-anniversary project changed in the following ways:

• Internally, it became an opportunity to rethink the meaning of their work.

• Storytelling was added to sales materials, providing conversation starters in business negotiations.

• It was utilized on the recruitment site as content that conveys the company's 'uniqueness.'

By communicating "who the company is and where it is heading" through its anniversary, Ishida was able to transform a simple internal event into the "core of corporate branding."

■ Summary of AI Tools and Prompts Used

Tool | Use Case | Prompt (Excerpt)
ChatGPT | Anniversary story design and structure | "Construct an anniversary story that conveys the inheritance and evolution of our philosophy."
Claude | Rewriting company history and founder episodes | "Make the episodes from the time of founding dramatic and relatable."
NotebookLM | Document analysis and keyword extraction from past to future | "Extract common keywords that verbalize our 'uniqueness' from past sales materials and recruitment brochures."
Canva | Visualizing timelines and vision materials | "Create a timeline design that conveys the company's vision and historical transition at a glance."
Genspark | Outputting templates for anniversary content design | "Generate a content structure template for an anniversary special website."

■ Column: A "Milestone" is a Trailer for the Next Story

An anniversary is not just an event to look back on the past.

It is the moment to redefine what the company "has believed in until now" and "where it is going from here."

Generative AI becomes a tool to turn the "material" of company history into a story and translate philosophy into the future.

It is not made to celebrate, but made to believe in the journey and tell the future.
That is how to use an anniversary as a PR professional.

■ What to do in these situations?

Q: I read the company history but it doesn't resonate. Where should I start?
→ Use NotebookLM to analyze the founder's words and corporate philosophy, and use ChatGPT to convert them into a "story worth telling."

Q: Employee enthusiasm is low and I can't get their cooperation.
→ Use Claude to organize it into "small stories where employees are the protagonists," and use Notion to template the interview design.

Q: The anniversary content feels like it will be too "insider-focused."
→ Use Genspark to generate multiple "content ideas for external audiences" and use Canva to translate them into visuals to achieve balance.

■ Reader Workshop: "Anniversary Story Design Workshop"

1. Use NotebookLM to read and analyze company history, representative messages, brochures, etc.

2. Use ChatGPT to reconstruct "the company's past and future" into a story format.

3. Use Claude to refine it into a soft and easy-to-read tone.

4. Use Canva to design timelines and vision visuals.

5. Use Genspark to create templates for two axes: "for external audiences" and "for internal audiences."

■ Finally

An anniversary is not a day to decorate past glories, but a day to polish words for the future.

The role of public relations is to articulate the meaning of a journey and give persuasive power to the next step.
By walking alongside generative AI, a company's "milestones" transform into "stories that remain in someone's heart."

29 | How to communicate public-private partnership projects?

(Chapter 3 | Building Relationships and Empathy - Dialogue Skills with Internal and External Stakeholders)

■ Introduction: How to bridge the "gap in intentions" between government and companies?

Edlinks, an education-focused startup, launched a career education project for high school students in collaboration with the local city of Fujigaoka.

Public relations representative Yuya Shibata had held numerous meetings with the city office staff, but when it came time to start the PR campaign, he felt hesitant.

• "The city wants it to look 'official'."

• "But we want to convey it as an 'innovative challenge'."

• "I don't know which narrative should be the focus..."

Both the press release drafts and social media posts ended up feeling "wishy-washy," and the intended message was not getting across.

■ Case: Finding common "words of intent"

Shibata first organized his previous exchanges using NotebookLM. He uploaded the city's policy documents, internal project proposals, and meeting minutes to analyze the trends and temperature differences in their language.

Prompt: "Compare the following multiple documents and analyze the differences in 'tone' and 'priority points' between the government side and the corporate side. Also, extract keywords of 'intent' that could be common to both sides."

As a result, shared values such as "wanting to nurture future leaders" and "wanting to create connections with the region" emerged.

Next, he input the following prompt into ChatGPT:

Prompt: "Regarding the public-private partnership project, please create a PR text that can be told as 'one story' while valuing both perspectives. Please structure it and set a tone that all stakeholders can be proud of."

Furthermore, he adjusted the text for the government using Claude, illustrated the key points with Copilot, created visual materials with
Canva, and structured a "Public-Private Partnership Project Communication Template" with Genspark.

■ Solution: Make the "purpose" the subject, not the "position"

Ultimately, Shibata switched to designing the PR starting from the "purpose of the project" rather than from the respective "government" or "corporate" positions.

• For the media, he generated buzz in the context of "Education x Regional Revitalization"

• Introduced in the city's public relations magazine as a case study of "co-creation between citizens and companies"

• Serialized in the company's own media as a record of "challenge and trust"

As a result, the relationship with the local government strengthened, and inquiries from other municipalities began to come in.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt)
NotebookLM | Analysis of documents from both government and companies | "Extract common values from multiple policy documents."
ChatGPT | Designing a story structure that connects both parties | "Transform the collaboration between the government and the company into a single story."
Claude | Rewriting into a public tone | "Adjust the tone to be easy to read and polite for government officials."
Copilot | Diagramming and creating slides for key points | "Express the purpose and flow of this project in a single diagram." Canva | Creating visuals that serve as a common language | "Design public relations materials that convey the relationship between citizens and companies."
Genspark | Generating public relations templates for public-private partnership projects | "Create a basic structure template for communicating public-private partnership initiatives."

■ Column: PR for "co-creation" begins with translating "standpoints"

Even when talking about the same thing, the government and companies use different "words".

It is the power of public relations that bridges that gap.

Generative AI can become that "translator of words."

While fine-tuning the tone and structure, you can design messages that depict a "shared future."
Let's cultivate "words that connect purposes" together with AI, rather than focusing on organizational convenience.

■ What to do in these situations?

Q: I'm having trouble because the tone used by the government and the company is too different.
→ Use NotebookLM to compare and organize documents from both sides, and use Claude to adjust them to an intermediate tone to make it softer.

Q: The significance of the project is "vague" and not being conveyed.
→ Ask ChatGPT to propose a "story structure that stakeholders can be proud of."

Q: I cannot maintain consistency between materials for the government and materials for the media.
→ Create communication templates for different purposes using Genspark, and diagram them using Copilot to make them easier to reuse.

■ Reader Exercise: Design exercise for words that "convey beyond standpoints"

1. Upload government and company documents to NotebookLM for analysis

2. Output a "PR structure plan based on a common purpose" using ChatGPT

3. Adjust to a government-oriented tone with Claude, and create a structured diagram with Copilot

4. Create visuals that serve as a common language using Canva

5. Save a "public-private partnership communication template" using Genspark

■ In Conclusion

When different perspectives meet, it is only natural for words to be misunderstood.

That is precisely why the role of public relations is to become an "interpreter of words."

Generative AI supports that translation work and helps refine words for the future.
Public relations, which connects the public and private sectors, is a "job of building bridges of meaning."

30 | PR not trusted by the company... Rebuilding internal branding

(Chapter 3 | Relationship Building and Empathy - Dialogue Skills with Internal and External Stakeholders)

■ Introduction: Who is internal PR actually reaching?

IT solutions company "Actways." Approximately 500 employees.

Mizuki SatoMizuki Sato from the PR department has been in internal PR for one year after transferring from the sales department. Although they produce an internal newsletter every month, they once overheard a conversation between colleagues.

It hit them hard.
It was true; there was little reaction to the content they were putting out. Feedback from employees was almost non-existent.

• "Is it becoming a one-way street?"

• "Is it not 'designed' to be read?"

• "Is it really getting through?"

Sato began to feel the need to rethink internal branding.

■ Case: Rethinking language and structure based on how employees 'receive' information

First, Sato usedPerplexity to collect examples of "well-read internal newsletters."
Next, they uploaded past internal newsletters, survey results, and Slack employee posts toNotebookLM and analyzed them using the following prompt.

Prompt: "Analyze the following internal communication-related information and extract 'themes and expressions that employees empathize with or react to.' Also, visualize areas for improvement and the 'temperature of the reader.'"

→ The analysis results revealed a tendency for reactions to gather around "articles where employees' faces are visible" and "real voices from the field."

So, they askedChatGPT the following:

Prompt: "Please restructure the following internal newsletter content so that employees can easily relate to it. Design it with an easy-to-read tone so that stories from the field and insights from work are conveyed effectively."

In addition, I adjusted the tone to be "not pushy" using Claude. I used
Notion AI to design the internal newsletter categories and organize tags. I designed a "composition for being read" using
Canva and created an "inner branding prompt template" using Genspark.

■ Solution: Redesigning the "structure of being read" to generate empathy

Sato realized that what is needed for internal public relations is not just the "power to convey," but a "structure to be received."

• Can readers think the titles and headings are "relevant to me"?

• Does the expression align with the "internal atmosphere"?

• Is the content "designed to be read within work hours"?

After scrutinizing these points, he proceeded to design an "empathy flow" utilizing AI.

As a result, the internal newsletter's readership rate increased by 42% compared to the previous month. The number of shares on Slack also increased, and voices from other departments began to arrive, saying, "I want you to introduce this in our department too."

■ Summary of AI tools and prompts used

Tool Usage Scenario Prompt Used (Excerpt) Perplexity Research on trends in other companies' internal newsletters and inner branding "Please tell me about successful examples of internal newsletters that are read and the common points in their composition." NotebookLM Extraction of employee reaction trends and areas for improvement "Please visualize the trends in internal expressions that are resonating and expressions that should be improved." ChatGPT Empathy design and storytelling for internal newsletter articles "Please re-edit this into a structure that conveys the real work and insights of employees." Claude Tone adjustment (soft, considerate expressions for the reader) "Please make this into a natural internal PR text that does not feel pushy." Notion AI Information organization and tag structure for internal newsletters "Please redesign the internal newsletter categories and organize them into an easy-to-read tag system." Canva Visual design and heading design "Please create a visual format that turns this content into an 'internal newsletter that gets read'." Genspark Creating prompt templates for inner branding "Please create composition templates for strengthening internal PR, categorized by purpose."

Column: An "internal newsletter that gets read" is one where "empathy begins before reading"

Before the PR team speaks, employees are already judging whether they "feel like reading it."

That is precisely why a "structure that makes people want to read" is necessary, rather than a "structure that forces them to read."

Generative AI becomes a partner that stands in the reader's shoes and designs the composition, tone, and flow starting from empathy.

Rather than the technique of conveying, it is about the design of being close to the reader. That is the beginning of trusted internal public relations.

What should I do in these situations?

Q: The internal newsletter feels like "company propaganda" and isn't read.
→ Empathy increases when you restructure it with ChatGPT to "make employee stories the protagonist" and soften the tone with Claude.

Q: The content of the articles is scattered and hard to read.
→ Organize categories with Notion, and grasp and unify composition trends with NotebookLM.

Q: I cannot measure whether it was read or what the reaction was.
→ While giving visuals a sense of guidance with Canva, design a "sharing flow" using Slack or internal tools.

■ Reader Workshop: What kind of 'face' does your internal communication have when it reaches employees?

1. Upload past internal newsletters, internal posts, and survey results to NotebookLM to analyze reaction trends

2. Rewrite into an 'empathy-driven structure' using ChatGPT

3. Re-edit into a gentle tone using Claude

4. Redesign categories and tags in Notion

5. Adjust the design for 'readability' in Canva

6. Create templates in Genspark to ensure consistency

■ Conclusion

Internal communication should not be the 'voice of the company,' but the 'voice that connects employees and the company.'

It might not be that they aren't reading it, but that the 'structure makes it hard to want to read.'

By using generative AI, you can refine the thoughts you want to convey into a 'form that reaches them.'
Being read is the first step toward being trusted.
That is where internal branding begins.

31 | How to communicate a change in leadership? Techniques for designing messages that 'move people'

(Chapter 4 | Acquiring 'communication weapons' you can use in the field)

■ Introduction: Vague handovers, stagnant teams

'Mirai Koki,' a group company of a long-established manufacturer.
Established 30 years ago with 200 employees. A few years ago, the management structure was revamped, and a new president from a venture background took office.

However, this kind of atmosphere was drifting through the company.

• 'When will things change?'

• 'Who is deciding what?'

• 'What is the direction of the new management structure, anyway?'

Public relations representative Satoshi Onuma was tasked with creating an internal message announcing the change in president, but he was bewildered.

■ Case: Structuring in the order of 'Reassurance, Expectation, Future'

Onuma first used ChatGPT to generate multiple patterns of 'internal messages for a change in president' to identify structural trends.

Next, he input past statements and internal newsletter comments from the outgoing and incoming presidents into Claude to extract the 'differences in language' regarding their speaking styles and visions.

Prompt example (ChatGPT): 'Please create an internal message to inform employees about a change in president at a manufacturing group company. Structure it in the order of "Reassurance, Expectation, Future," and write it in a way that the staff can accept positively.'

He organized the extracted keywords and tone, and reconstructed it in the following flow:

1. [Reassurance]: Background of the change and gratitude to the outgoing president

2. [Expectation]: The personality and mindset of the new president

3. [Future]: The direction we are heading, what will change, and what will remain the same

Canva was also used to attach a 'Vision Map' with diagrams, finishing it with a structure that connects emotion and understanding.

■ Solution: It is the structure and order that changes the atmosphere

What employees were looking for was not 'positive feelings' but an 'order that provides reassurance'.

• Being told about philosophy suddenly does not resonate

• Slogans alone cannot be felt as reality

• Communicating who, why, and how things will change in order leads to trust

As a result, the internal survey after the announcement received many comments such as 'It was convincing' and 'I could see the direction for the workplace'.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt) | ChatGPT | Generation of internal message structure | 'Please create a message announcing the change of president to employees in the order of "Reassurance, Expectation, Future"' | Claude | Extraction of speaking tone and linguistic alignment of old and new presidents | 'Summarize the differences in speaking style and values of the two people below, and adjust them to a tone that reaches the employees' | Canva | Creation of internal visuals | 'Please create a map that visually conveys the vision of the new system after the change'

■ Column: For people to 'understand and act', order is 90% of the battle

In times of change, 'how you communicate' becomes more important than the information itself.

Especially at milestones like a change in leadership, if the employees' hearts are not moved first, the practical work will not follow.

That is precisely why,
a design that changes the 'atmosphere' in the order of 'Facts → Gratitude → Expectations → Will' is necessary.

AI becomes a powerful editorial partner for rearranging and refining the thoughts you want to convey from the 'reader's perspective'.

■ What to do in these situations?

Q: The new president's words are stiff and hard to relate to...
→ Input the speech notes into Claude and have it rewritten in a 'tone that reaches the employees on the front lines'.

Q: The message gets blurred when focusing only on the fact of the change
→ Ask ChatGPT for a 'story structure that can be received from the employees' perspective'.

Q: When 'who says it' is important, how should PR design it?
→ Be conscious of 'designing to show the speaker and the content as a set', such as using Canva for 'vision maps' or 'excerpts of the inauguration message'.

■ Reader Workshop: Turn 'Change Message Design' into an Internal Template

1. Create multiple structural drafts of 'Reassurance → Expectations → Future' using ChatGPT

2. Refine the speaking style of the new leadership team using Claude

3. Diagram and visualize the 'direction of change' using Canva

4. Save as a 'Change Message Template' that can be shared within the company

■ In Conclusion

A change in president is the moment the company's 'face' changes.
It is not just a fact, but also the beginning of a 'story heading toward the future'.

How you tell that story will greatly change the atmosphere within the company.
Generative AI is a bridge that connects the president's 'will' with the employees' 'feelings'.

'Structural power that does not fear change' is exactly the situation where the true value of PR is tested.

32 | Seminar presentation materials: What and how should you speak to make an impact?

(Chapter 4 | Acquiring 'Communication Weapons' You Can Use in the Field)

■ Introduction: The reason for the frustration of 'The content was good, but it didn't resonate'

Treasure Hub, a B2B marketing support company.
Public relations manager Yuto Okamoto recently supported the company president's seminar presentation materials.

The slides were polished. The track record was strong. The content should have been solid.
However, the audience's reaction was lukewarm.

• Survey comments were "average" and "felt like I've seen this before."

• There was no buzz on social media.

• There were no requests for future speaking engagements.

Okamoto decided to rethink the "compelling structure" itself.

■ Case: Changing to a structure that tells a "Before -> Change -> Story" narrative

First, Okamoto gave ChatGPT the following prompt.

Prompt: "Please redesign the following seminar content into a 'memorable structure.' Create a narrative arc for the entire presentation and focus on a delivery style that 'hooks the audience within the first 10 minutes.' Also, please insert a 'before-and-after episode that resonates with the listeners.'"

Furthermore, he had Claude read past presentation materials and scripts to analyze "redundant parts," "points that should be emphasized," and "structures that remain in the listeners' memory."

As a supplement, he uploaded past seminar evaluations and surveys to NotebookLM to extract commonalities between "elements with high engagement" and "elements that didn't resonate."

For the finishing touch, he used Canva to create a visual slide that would be "the one slide to remember."
He structured it to pair words that deeply resonate with the story with a powerful visual.

■ Solution: Refining the structure by working backward from "why we are telling this"

What Okamoto realized was that it is the "structure and order," not the "content of the talk," that determines the reaction.

• "Conclusion first" alone is not enough to resonate.

• "Spec explanations" alone do not create empathy.

• Only with a "story of change" will people be convinced.

As a result, satisfaction levels in the survey for the next presentation improved significantly.
Comments that led to action, such as "I was drawn into the talk" and "I wanted to implement this at my own company," were gathered.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Usage Prompt (Excerpt)
ChatGPT | Redesigning story structure and tone | "Please redesign this into a memorable seminar structure. Create a narrative arc with a focus on a 'compelling 10 minutes'."
Claude | Script analysis and extraction of key points | "Please refine this presentation draft into a 'memorable structure'. Point out parts that should be cut and parts that should be emphasized." | NotebookLM | Analysis of past evaluation trends | "Compare and analyze patterns with high and low ratings based on the uploaded seminar materials and surveys."
Canva | Creating impact slides | "Please turn this into a visual that makes the 'decisive slide' of the presentation memorable in a single image."

■ Column: The 'depth' of a slide is determined by the starting point of the words

Common issues with seminar materials are 'overloading' and 'over-explaining'.

It is not about packing in information, but rather designing the flow of words and the pauses that creates understanding and empathy in the listener.

Generative AI is not just for paraphrasing, but an editing partner for organizing the 'order of the narrative'.

Polishing a slide begins with 'finding the starting point of the narrative'.

■ What to do in these situations?

Q: The president's talk is too abstract, and I'm having trouble with the structure
→ Let's have ChatGPT provide multiple 'story-based structure proposals'.

Q: The content is good, but the reaction is weak...
→ It is effective to have Claude analyze past presentation drafts and point out 'parts that lack emphasis' or 'parts where the development is weak'.

Q: I can't create a memorable slide
→ Let's convey the 'main argument' of the presentation to Canva and have it 'visualize the decisive slide'.

■ Reader Workshop: Preparing a 'Seminar Material Template' for your company

1. Create a 'narrative arc + before-and-after' structure proposal with ChatGPT

2. Polish scripts and slides into 'words that strike a chord' with Claude

3. Analyze reaction patterns by cross-referencing past surveys and evaluations with NotebookLM

4. Create a memorable 'decisive slide' with Canva

■ Conclusion

PR needs 'stories that move people' rather than just 'good stories'.
To achieve this, the key is not just what you say, but 'how you structure it'.

Generative AI is an entity that connects the speaker's intent with the listener's expectations, creating the axis of the structure.
You can rely more on AI for seminar design that generates empathy and action.

33 | How to get a president's contribution published? From media selection to editorial handling

(Chapter 4 | Acquiring 'communication weapons' that can be used in the field)

■ Introduction: "I have something to say, but I can't write it down"

SaaS startup "SelectBiz." 8 years since founding, 80 employees.
Public relations representative Shota Matsushita received a request from the CEO.

Matsushita immediately created a proposal, but the CEO's feedback was vague.

• "I want to put more 'passion' into it"

• "But it shouldn't be too casual either"

• "Is this media a bit too stiff?"

As a result, both the CEO and the PR team became lost regarding what to write, for whom, and how.
Because the CEO is busy, the alignment process didn't progress as expected.
"It's harder to 'get it approved' than it is to 'write it'"—that was Matsushita's honest realization.

■ Case: Using AI to balance "what you want to say" with "a structure that gets approved"

First, Matsushita used Perplexity to organize a list of media outlets that accept contributed articles and their respective characteristics.
He used this as material to select which media would be a good fit.

Next, he entered the following prompt into ChatGPT.

Prompt: "Based on the following notes from the CEO's remarks, please create a 'structure that will be accepted as a contributed article.' Please adjust the tone of the language and the inclusion of examples to match the media's editorial stance, and conclude the piece with a 'managerial perspective.'"

Furthermore, he had Claude read past CEO interviews and speech drafts to extract a writing tone that conveys 'this person's personality.'
He aimed for a finish that sounded like the person's own words while polishing the content to look professional.

Once the draft was compiled, he used NotebookLM to cross-reference it with internal company documents to ensure there were no discrepancies in the statements.

■ Solution: Designing a bridge between "the CEO's words" and "the reader's perspective"

Matsushita realized that the reason "the CEO cannot write" is not a lack of "literary talent," but a lack of "visibility into the structure."

• Not knowing how to start writing

• Difficulty judging what tone of voice suits the media's readers

• Thoughts in the head not turning into 'polished text'

That is precisely why the role of PR is not to "polish," but to "connect."

As a result, the manuscript was published successfully, and the article was widely cited on X (formerly Twitter).
The media outlet even reached out, saying, "We'd like to work with you on other topics as well."

■ Summary of AI Tools and Prompts Used

Tool | Use Case | Prompt (Excerpt)
Perplexity | Collecting media characteristics and trends | "List 10 media outlets that frequently feature contributions from startup executives, categorized by genre."
ChatGPT | Manuscript structure and tone adjustment | "Based on these notes, create a contribution structure that resonates with the readers and the media outlet. Conclude with a summary from a management perspective."
Claude | Reproducing the CEO's tone and style | "Extract the consistent 'voice' from past speeches and reflect it in the text."
NotebookLM | Checking consistency between the manuscript and internal information | "Check if there are any contradictions between the following manuscript and internal documents, and point them out if necessary."

■ Column: A contribution begins the moment you decide 'who you are talking to'

A contribution is a valuable opportunity to deliver the CEO's 'voice'.
However, if the audience is unclear, the words become increasingly blurred.

If you are clear about 'who you want to reach and what you want to convey,' both the content and tone will fall into place.

Generative AI is a partner that helps clarify that 'reader persona' and gives shape to the narrative.

When you can visualize the person you want to reach, the contribution finally becomes 'words'.

■ What to do in these situations?

Q: The CEO's words become too stiff when turned into a manuscript.
→ Use Claude to extract the tone from the 'original speech' and ChatGPT to restructure it.

Q: I can't decide which media outlet to submit to.
→ The fastest way is to use Perplexity to identify 'contribution track record vs. media trends'.

Q: It's difficult to coordinate with a busy CEO...
→ Compile past statements and documents in NotebookLM, and the AI will organize the 'intent' for you.

■ Reader Workshop: Creating a 'Draft Contribution Structure' with AI

1. Generate multiple 'contribution structure' drafts from speech notes using ChatGPT

2. Refine the text to reflect the 'CEO's personality' using Claude

3. Check for consistency with internal documents using NotebookLM

4. Confirm compatibility with the media outlet using Perplexity and submit the proposal

■ Conclusion

A CEO contribution is a valuable venue where the face of the company speaks to society.

Whether that voice can be shaped to truly 'reach' the audience.
That is where the PR professional's design capability is tested.

By utilizing generative AI, you can bridge the gap between words and structure, crafting manuscripts that people will actually read.
Let's work together with PR and AI to build a bridge that connects the 'narrative in the president's mind' with society.

34 | Book Publishing as a PR Strategy: From Planning to Generating Buzz

(Chapter 4 | Acquiring 'Communication Weapons' You Can Use in the Field)

■ Introduction: 'If we're going to publish, I want it to be a book that spreads'

ArchiLink, a community-based real estate tech company. 12 years since establishment, 120 employees.

The company's PR manager, Mika Ishida, was consulted by the management team as follows.

A book publishing project had already been launched as a 10th-anniversary initiative, but—

• 'It must not become just a company history.'

• 'But if it sells too well as a business book, it won't feel like us.'

• 'We don't know what the core of the plan should be.'

For a PR team with no publishing experience, this was a 'design from scratch' project.

■ Case: Using AI to identify the 'core narrative' and translate it into a structure

First, Ishida consolidated past internal documents, interview articles, and anniversary messages into NotebookLM. She used the following prompt to identify the 'common threads of corporate identity'.

Prompt: 'Read the following internal documents, interviews, and past anniversary greetings, and extract the 'common narrative threads' and 'consistently discussed themes' that are characteristic of this company.'

Then, she used ChatGPT to create a structural proposal as follows.

Prompt: 'Based on the 'common narrative threads' and 'corporate themes' extracted above, propose a 'story-driven structure' for planning a book as a PR strategy. The target audience is business professionals and company stakeholders.'

Furthermore, she created initial visual concepts in Canva to prepare internal sharing materials. She used
Claude to assist in proofreading the draft manuscript to ensure 'consistency and reader understanding.' She used
Perplexity to collect publishing examples from competitors and clarify the differentiation of the theme.

■ Solution: A book that balances the 'purpose of the narrative' with a 'structure that gets read'

What Ishida kept in mind until the end was bridging the gap between 'internal language' and 'how it is read externally.'

• Significance as a 'record' within the company

• Appeal as a 'message' to those outside the company

By being conscious of this balance, the final title, structure, and tone of voice all reflected the company's unique identity.

After publication, it was utilized in recruitment PR and corporate branding, becoming a book that lives on as a long-term PR asset.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Prompt (Excerpt) | NotebookLM | Analysis of internal documents and anniversary messages | 'Extract the commonalities of "corporate identity" from this set of documents' | ChatGPT | Creation of book structure proposal | 'Propose a structure that can be used for PR strategy, leveraging the company's narrative themes' | Claude | Consistency of manuscript and reader-friendly proofreading support | 'Check if this manuscript is "easy for external readers to read" and provide improvement suggestions' | Canva | Book cover design and creation of shared materials | 'Propose several visual images that fit this book theme' | Perplexity | Collection of similar cases and positioning analysis | 'Tell me about successful examples of company histories and business books published by small and medium-sized enterprises. Please include themes and media.'

■ Column: How to avoid 'over-narrating' when a company publishes a book

'I want to tell everything' is the first step toward a book that won't be read.
Whether it's a company history or a message book, if the subject is the company rather than the 'reader,' it won't resonate.

That is precisely why you should use generative AI to objectively view 'what might resonate.'
Rather than what you want to say, focus on 'how you want them to feel after reading it.'

A book as a PR strategy is a tool that works not for the 'past,' but for the 'future.'

■ What to do in these situations?

Q: I'm worried it will end up just being a company history
→ Use NotebookLM and ChatGPT to extract the 'common axes' that have been discussed and structure the reorganization.

Q: This is my first time publishing, where should I start?
→ You can grasp the flow by collecting successful cases with Perplexity and generating a 'book strategy template' with Genspark.

Q: The writing feels 'not like our company' while I'm drafting it
→ Get feedback on 'corporate narrative identity' from Claude.

■ Reader Workshop: Drafting a 'Company-Specific Book Structure' with AI

1. Have NotebookLM read past documents, speeches, and messages

2. Have ChatGPT generate three structural proposals that resonate with the target audience

3. Perform tone proofreading from the reader's perspective using Claude

4. Create initial cover drafts and slide materials in Canva for internal proposal

■ Conclusion

Book publishing can be the ultimate "PR material."

However, whether you can maximize that power depends on
whether you can create a structure that gets read and whether it becomes a point of contact with the outside world.

AI outlines the structure, extracts the essence of the narrative, and refines the tone.
It should become a powerful partner that transforms "corporate words" into a book that connects with society.

35 | Novelty distribution copy: What kind of words move hearts without sounding like a sales pitch?

(Chapter 4 | Acquiring "Communication Weapons" for Use in the Field)

■ Introduction: I want people to be happy. But I don't want to sound like I'm "selling"

Environmental startup "GreenRoot" adopted original bottles as a novelty item for an exhibition.
PR representative Takafumi Sakai was struggling with the draft for the message card to be distributed.

• What and how should I convey in a short message?

• How much should I talk about my own company?

• Is the writing style completely different from product introduction copy?

"PR-like, but warm."
He hadn't been able to find that "sweet spot."

■ Case: Placing the starting point of empathy on the "user's experience" rather than the "product"

Sakai uploaded his past exhibition feedback notes and background materials on the novelty production to NotebookLM.

Prompt: "From the following materials, organize the 'background of how this novelty was born' and 'the feelings we want to convey.' Also, propose an angle for the message that assumes the 'usage scene' of the user receiving it."

Next, he used ChatGPT to refine the writing style to "imagine the other person's feelings."

Prompt: "For an exhibition where we are distributing 'eco-friendly water bottles' as a novelty, please create a message that is warm and generates empathy. Instead of a company introduction, focus on 'what the person receiving this will feel.'"

Furthermore, he created the visual for the message card in Canva, and asked
Claude to check for "misinterpretations or awkwardness."

■ Solution: Talk not about the "novelty," but about the "experience beyond it"

What Sakai learned was to "change the subject of the words from the company to the 'other person'."

• Corporate initiatives → Changes in the user

• Talk of materials and structure → The scene of the future in their hands

• Message drafts → Converted into triggers for dialogue and empathy

As a result, the response at the exhibition was significant, and it generated a "spread" beyond imagination, such as being posted on social media with photos saying, "This novelty conveys the company's thoughts."

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Used Prompt (Excerpt)
NotebookLM | Organizing the background for creating distribution text and extracting assumed scenarios | "Summarize 'who this novelty will reach and in what kind of scene'"
ChatGPT | Creating distribution text with a warm tone | "Make it a short sentence that will make the recipient happy. Don't make it too 'corporate-focused'"
Claude | Text checking and preventing misinterpretation | "Point out if this text 'looks like a sales pitch' or 'could lead to misinterpretation,' and suggest improvements"
Canva | Visualizing message cards | "Create 3 warm and friendly card design ideas that fit eco-friendly novelties"

■ Column: The 'one line' of sales promotion and public relations are different

The purpose of sales promotion text is to 'convey.'
The purpose of public relations text is 'what remains after it is conveyed.'

Even with the same one line, the power of public relations is being able to 'make them feel.'
That is why, instead of just having AI rephrase it,
let's have it think together while imagining 'who it will reach.'

What you want to convey is not the novelty, but the 'feeling of thinking of you.'

■ What to do in these situations?

Q: The novelty looks like just a handout
→ Let's use ChatGPT to convert it into a narrative starting from the 'usage scene.' It is effective to change it from the structure.

Q: The way of speaking is inconsistent for each exhibition and lacks unity
→ Analyze past message texts with NotebookLM and design a common format with ChatGPT.

Q: People in the company say, 'I want to include more product introductions'
→ Get feedback from the reader's perspective with Claude and show an appropriate balance.

■ Reader Workshop: Redesigning the 'one line that the novelty speaks'

1. Have NotebookLM read internal documents, exhibition logs, and background documents

2. Create multiple 'short sentences starting from the other person's experience' with ChatGPT

3. Prepare it as a card in Canva and check the tone including how it is shown

4. Check with Claude whether the expression prevents misunderstanding

■ Finally

Novelties are the "gateway to words" that get people to pick them up.
That single message becomes a PR space that conveys a company's values.

Generative AI gives us the power to think from the "other person's perspective."
AI will be a reliable partner in transforming small touchpoints like handouts into words that cultivate trust.

36 | Recruitment Pitch Decks: How to Update to the Latest "Attraction Points"

(Chapter 4 | Acquiring "Communication Weapons" You Can Use in the Field)

■ Introduction: Pitch Decks That Are Attractive But Don't Get Through

B2B SaaS company "Efficion."
With 120 employees, they are actively hiring. Product growth is steady, but for some reason, the response to their "recruitment pitch deck" is poor.
Takuma Miyajima,Takuma Miyajima, who handles both PR and HR, was struggling with the results of post-interview surveys.

• "The deck feels a bit too pushy."

• "I wanted to know about the culture rather than the job description."

• "The design is outdated. I don't feel any momentum."

It was true; it hadn't been significantly updated since it was created three years ago.

■ Case: Extracting and Reconstructing Candidate "Attraction Points" with AI

First, Miyajima uploaded records of past interview surveys, exit interviews, and current employee interviews toNotebookLM.

Prompt: "Based on the following information, categorize the 'points that attracted people to Efficion' and the 'reasons for leaving.' Then, propose elements for a recruitment pitch deck that conveys 'what Efficion is like today.'"

Based on the output elements, he askedChatGPTfor a structure proposal and a draft storyline.

Prompt: "Create a recruitment pitch deck that makes candidates feel, 'I want to work at this company.' Re-edit it with an emotional structure, focusing on the following keywords."

For the design, he adjusted the template inCanva, and checked the phrasing of the narration text withClaude.

■ Solution: A Structure That Answers "Questions in the Heart" Rather Than Just a Deck

By reviewing not just the appearance but the structure itself, the following results were achieved.

• An increase in candidates who feel the "company culture comes through" and that they "can empathize."

• Materials anticipate questions, improving the quality of interviews

• Latest visual storytelling generates positive reactions on social media

What I was particularly conscious of was structuring the content around the subject of 'what the person I want to attract is struggling with'.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Prompt (Excerpt)
NotebookLM | Analyzing employee and candidate feedback | 'Classify the reasons for attraction and reasons for leaving, and extract the effective components for attraction'
ChatGPT | Recruitment material story structure | 'Create a recruitment pitch structure that resonates with candidates based on the following keywords'
Canva | Redesigning material visuals | 'Create visual template ideas for recruitment materials that would interest young candidates'
Claude | Refining expressions and narration | 'Review this draft material to see if it is easy to understand and not too emotional, and suggest improvements'

■ Column: You win when they think, 'You noticed my worries'

A good recruitment pitch is not about forcing information on someone.
It is about whether the candidate thinks, 'You said what I wanted to know first.'

Generative AI finds patterns from past voices and guides you to the 'structure required now.'
Materials should be something that speaks for the candidate's 'anxieties.'

■ What to do in these situations?

Q: The pitch material is too long and doesn't get read
→ Ask Claude, 'If you were to convey only the essence of the material verbally?' to extract the key points.

Q: Opinions are divided within the company on what to emphasize
→ Use NotebookLM to aggregate candidate voices and extract an objective 'common perspective' to serve as a starting point for discussion.

Q: I need to update it urgently but am not confident in the structure
→ Ask ChatGPT for a 'slide structure template + headline text' and use it in a state where it can be edited immediately.

■ Reader Workshop: 'Retelling the recruitment pitch material with the current company image'

1. Upload past interviews, conversations, and candidate notes to NotebookLM

2. Design a 'storyline that resonates with candidates' using ChatGPT

3. Refresh the presentation and update the visuals in Canva

4. Check and fine-tune the tone and impression of the language with Claude

■ Conclusion

A recruitment pitch material is not a 'record of past achievements,' but an 'invitation to empathy' for the future.
Only with a structure that is close to the heart of the person you want to attract will it become a message that truly moves them.

Generative AI is the partner that draws the blueprint.
AI holds the key to creating materials that leverage past voices and reach the future.

37 | Where do you start when rebuilding an old company profile?

(Chapter 4 | Acquiring 'Communication Weapons' for Practical Use)

■ Introduction: The discomfort of not being able to talk about the company's 'now'

Takara Seiko, a long-established manufacturing company. 45 years since its founding, with 250 employees.
Maki Tajima, in charge of public relations, was tasked with refreshing the company profile distributed at exhibitions.

However, anxiety grew from the moment she opened the document.

• 'The font is outdated'

• 'What it says hasn't changed in 20 years'

• 'Actual strengths and new initiatives are not reflected at all'

■ Case: Re-examining old materials from the 'skeletal structure of words'

Tajima first uploaded past company profile materials and brochures to NotebookLM.

Prompt: From the following company profile materials, extract 'expressions that feel outdated' and 'parts where strengths are buried,' and propose a structure that aligns with the current corporate image.

Next, she input the current materials into ChatGPT and consulted on 'how to reconstruct it as a modern company guide for exhibitions.'

Prompt: Please re-edit the following company introduction text as a 'corporate image that resonates with today's society.' To ensure it also resonates with young employees, please include connections to social issues, the significance of working, and challenges for the future.

As a finishing touch, she refreshed the visual templates in Canva and checked the politeness of the phrasing with Claude.

■ Solution: Updating materials becomes a 'reconfirmation of the organization'

Through the reconstruction process, a dialogue naturally emerged between Tajima and the management team.

• 'Where are we focusing our efforts now?'

• 'Changes in corporate culture and work styles'

• 'What we want to achieve in the next 10 years'

As a result, the materials that were supposed to be for "external explanation" transformed into "words for self-awareness" for the entire organization.

■ Summary of AI Tools Used and Prompts

Tool | Use Case | Prompt (Excerpt)
NotebookLM | Extracting points of obsolescence from past materials | "Please separate the outdated parts of this material from the elements that are still relevant today and present them."
ChatGPT | Re-editing and generating structural proposals | "Please restructure this company introduction from a modern perspective, making it a text that is conscious of its connection to society."
Canva | Refreshing visuals | "Please redesign the materials using modern templates without compromising the brand feel of the manufacturing industry."
Claude | Reviewing tone and politeness of language | "If there are any excessive expressions, outdated phrasing, or unclear parts in this text, please point them out and provide suggestions for improvement."

■ Column: "Inventorying" materials is also a "re-confirmation" of the company

Old company materials are not just relics of the past.
They serve as clues to highlight "what should continue to be told and what should be changed."

By utilizing generative AI, you can review "expressions that were repeated unconsciously."
It is closer to "recognizing our current position" than just updating words.

■ What to do in these situations?

Q: I'm in trouble because management says, 'It's fine as it is.'
→ Have Claude create a comparison text between the 'impression of the old materials' and the 'differences in the proposed new materials,' and show it to the management team.

Q: It's hard to draw the line between what is old and what is valuable.
→ Upload past materials to NotebookLM and have ChatGPT perform a 'modern perspective editorial evaluation.'

Q: I can't refresh the visuals well.
→ Choose a category like 'For Manufacturing/For Exhibitions' from Canva templates and adjust it to your brand colors.

■ Reader Workshop: 'Value Inventory Workshop for Company Introduction Materials'

1. Upload the last 5 years of materials to NotebookLM

2. Extract elements of a 'corporate image that resonates with today's society' using ChatGPT

3. Redefine the message through dialogue with the management team

4. Update visuals in Canva and adjust expressions with Claude

■ Conclusion

Materials are not just a 'company introduction,' but a symbol of 'what the company values.'
Generative AI becomes a translator that connects past accumulation with future will.
Re-answering the question 'Who are we?' through the process of creating materials.
That is the true meaning of company introduction materials.

38 | PR Strategy and How to Polish Applications to 'Win' Awards

(Chapter 4 | Acquiring 'Communication Weapons' You Can Use in the Field)

■ Introduction: 'I applied, but I didn't get a response'

SaaS startup 'CrevoLink'. In its fourth year of business.
The person in charge of PR, Yosuke Nakai, was told by the company, 'It's time we aim for an award,' and sent applications to several startup awards.

However, the results were not favorable.

• Rejected at the document screening stage

• Not particularly picked up by the media either

• Pressed by management, 'What was missing?'

■ Case: Learned that awards are a 'game of reading the context'

Nakai first conducted thorough research on 'past award-winning companies' and 'trends in application content' for the same awards using Perplexity.

Next, he uploaded past application texts, PR materials, and media articles to NotebookLM. He realized that past applications 'only emphasized features and achievements, and lacked the context of "why this company, why now?"'

ChatGPT was given the following prompt.

Prompt: Read the award application text below and rewrite it into an 'award-winning application' by focusing on: - A structure that evaluators want to read - The connection between social background and the business - The impact worthy of an award.

Furthermore, he used Claude to check and refine the tone, and Canva to prepare illustrative materials for the application.

■ Solution: Working backward from 'who you want to be evaluated by and for what reasons'

Instead of 'writing what we want to say,' he reconstructed the language design based on 'how it fits what this award wants to evaluate.'

• Connection to social background

• Talking about the 'future,' not just numbers

• Not how we are different from other companies, but 'the reason we shine in this context'

Based on this perspective, he organized an 'award application prompt collection' using Genspark. He was now able to share the perspective on which awards to target with management.

■ Summary of AI tools used and prompts

Tool Usage Scenario Used Prompt (Excerpt) Perplexity Research on award trends and context 'Please tell me the past winners of this award and the reasons they were chosen.' NotebookLM Comparative analysis of past applications and PR documents 'Extract a consistent message from application texts, PR materials, and articles, and organize the elements to be used in the application.' ChatGPT Restructuring and adding impact to application text 'Reconstruct with a structure and story that passes the award, and emphasize social significance.' Claude Refining tone and wording 'Please adjust this text to a tone that is easy for award judges to read and gain empathy.' Canva Illustration and visualization of materials 'Please create illustrations that fit the application. Use a structure that integrates the message and visuals.' Genspark Organizing prompt templates 'Please organize prompt templates for award applications by industry and purpose.'

■ Column: Awards are about telling a 'story that gets chosen'

Award judging is not just an evaluation.
It is a battle of whether you can fully articulate 'why this project is needed in society right now.'

Features and achievements are merely the foundation.
Without a 'story' on top of that—explaining 'why this company is here, right now'—you cannot move people's hearts.

Generative AI provides thorough support for 'verbalizing to be chosen.'
What is needed is not polishing words, but creating a structure that conveys your intent.

■ What to do in these situations?

Q: The evaluation criteria for award applications are unclear, and I cannot prepare.
→ Use Perplexity to collect past winning comments, and use NotebookLM to compare elements and extract commonalities.

Q: I was told my application lacks persuasiveness.
→ Use ChatGPT to reconstruct a story that reinforces the 'connection to social issues' and 'future perspective.'

Q: My application materials are just text and are boring.
→ Create impactful diagrams with Canva and supplement them with visuals.

■ Reader Workshop: Designing an 'award-winning story structure' for your company

1. Understand the winning trends of the target award using Perplexity

2. Analyze your company's past PR and application texts using NotebookLM

3. Reconstruct as a 'context-aligned structure' using ChatGPT

4. Add visuals with Canva and refine the expression with Claude

5. Accumulate an 'application template collection' using Genspark

■ Conclusion

Applying for an award is about redefining your company's value.
In the process of searching for the words to be chosen, you can also return to your origin: 'Why are we doing this?'

Generative AI can be your partner in that dialogue.
The process of verbalizing not just the 'result' of winning, but the 'meaning' leading up to it, becomes the greatest asset for public relations.

39 | PR techniques for balancing attendance and media coverage for company-hosted PR events

(Chapter 4 | Acquiring 'communication weapons' you can use in the field)

■ Introduction: 'Who is this event for?' was ambiguous

Axoltech, a long-established IT company with 500 employees.
To commemorate the 10th anniversary of their product, they decided to hold a live event.

Public relations representative Soichiro Yoshida was tasked with designing this event, which was expected to be a topic of conversation both inside and outside the company, but preparations were struggling.

• “We are focusing too much on attracting attendees, and the message we want to convey is becoming diluted.”

• “Preparation for press relations is falling behind.”

• “The speakers' talks are inconsistent regarding who we want to reach.”

■ Case: Changing the subject defined the core of the project

Yoshida first consulted ChatGPT. To articulate the purpose and message of the event, he used the following prompt.

Prompt: “Based on the overview of the 10th-anniversary event and internal feedback, please propose a restructuring of the project from two perspectives: ‘how we want attendees to remember it’ and ‘how we want the media to report it.’”

→ Result: It turned out that a story about “how we have contributed to society over the past 10 years” was more likely to be picked up by the press than “the future of the product.”

Claude was used to structure the speakers' speeches and organize key points.
Perplexity was used to research commemorative events and reporting trends of competitors.
Canva was used to prepare visual materials for event announcements, and Notion was used to centralize internal communication and preparation tasks.

■ Solution: The design of “Experience x Media Story” is vital for an event

The key points of the project that Yoshida restructured are as follows:

• For attendees: Clarify the “experiential value” (demo experiences, dialogue corners, etc.)

• For the press: Design a “symbolic message that is easy to put into words” (connection to social context)

• The speakers' words were also tuned to make them easy to use as “article headlines.”

Thanks to this design, the event received the highest number of press mentions ever. The attendee satisfaction survey also received positive feedback.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Used Prompt (Excerpt) | ChatGPT | Event design and purpose organization | “To make the event memorable, what and how should we design it from both the attendee and press perspectives?” | Claude | Speech structure and extraction of key points | “Please reorganize the speakers' talks into a structure that is easy to turn into an article.” | Perplexity | Research on competitor events and reporting trends | “When commemorative events in this field were reported, what elements were picked up as news?” | Canva | Visualization of announcement materials and presentation slides | “Please create flyer and material visuals that fit this event announcement.” | Notion | Internal coordination and task management for day-of operations | “Please list the preparation tasks required for this event.”

■ Column: Design the event as a “place to be talked about”

Events often become an end in themselves, where the goal is just to 'get it done'.
However, the true skill of a PR professional lies in how they create a 'reason to be talked about'.

What is worthy of news coverage is not just a simple event.
'Why was this event created now?' and 'What do we want to share with society?'
Articulating these points and refining them to be easily communicated is the core of utilizing AI.

■ What to do in these situations?

Q: The event feels like it's just for internal staff.
→ Have ChatGPT diagnose 'what needs to be fixed to make it suitable for an external audience'.

Q: The speaker's talk is too long and the main points are unclear.
→ Use Claude to 'summarize it into three messages and turn them into catchphrases' to organize the content.

Q: There isn't enough evidence to include in the press release.
→ Use Perplexity to pull industry trends and public opinion surveys to supplement data that connects to the social context.

■ Reader Workshop: Designing the balance between 'Attracting Attendees' and 'Media Coverage' for your company event

1. Use ChatGPT to flesh out event designs from both the 'attendee perspective' and the 'media perspective'.

2. Use Claude to restructure speaker scripts and organize key points.

3. Use Canva to create unified visuals for invitations and presentation materials.

4. Use Notion to centrally manage preparation tasks and progress.

5. Use Perplexity to compare media coverage examples with your theme to determine the angle for coverage.

■ Conclusion

An event is the 'greatest real-world media' for PR.
However, its value is determined not by 'doing it', but by 'how it remains in memory'.

By utilizing generative AI, you can proactively design a structure that will be remembered, talked about, and covered by the media.
That is what leads to the true results of a PR event.

40 | Design skills to overcome the 'Who is the subject?' problem in collaboration projects

(Chapter 4 | Acquiring 'Communication Weapons' for the Field)

■ Introduction: The temperature difference that seeps through despite calling it a 'Joint Announcement'

TechField is a SaaS company that develops services in the B2B sector.
One day, a collaboration project with the industry's largest company was decided.

PR representativeShunta Muraiwas entrusted with the production of a joint press release and campaign page.
However, he hit a wall early on.

• Clashing over "which company name should come first"

• The submitted drafts are both "self-promoting" and fail to align

• With every agreement, the text becomes increasingly mediocre

■ Case: Changing the "subject" of the text to the customer

Murai first asked ChatGPT to organize a third option based on the strengths of both companies and customer benefits. The following prompt was used.

Prompt: "Analyze the following press release drafts from two companies and reconstruct them into a form that looks attractive from the reader's perspective. Express the strengths of both companies as 'customer value' and refine the writing style so that it does not lean toward either side."

→ What ChatGPT proposed was a structure that "does not make the company names the subject."
It was a third way of speaking that started with "What changes with this partnership?"

Next, Claude was used to analyze the partner company's past releases and extract differences in tone and writing style.
Genspark was used to organize the characteristics and role distribution of both companies as a "collaboration text template."

Visual materials were designed in Canva to maintain a sense of "equality" and "consistency."

■ Solution: Balancing by "making the reader the subject"

The solution Murai arrived at was to stop asking "which one is the subject?" and instead use a structure where "what the reader receives" is the subject.

• Replace the subject from "company" to "value" or "change"

• By making the starting point of the message the "customer's situation," it is easier to gain empathy

• The "equality" of both companies is ensured through tone and design, not by pushing it in the wording

Through this approach, consensus-building between the two companies proceeded smoothly, and the press materials were refined into impactful text.

■ Summary of AI tools and prompts used

Tool Usage Scenario Prompt Used (Excerpt)
ChatGPT Reconstructing drafts from both companies from a reader's perspective "Please create a press release draft that is neutral, customer-value-centered, and does not lean toward either side."
Claude Analyzing the partner company's past tone "Extract the characteristics of this company's release style and structure, and analyze how they differ from our own."
Genspark Designing collaboration text templates "Organize a 'collaboration text template' that can be used between two companies by element."
Canva Designing consistency for visual materials "Propose slide or material designs that balance and integrate the logos, colors, and elements of both companies."

■ Column: The "design power" of collaboration text is about dialogue and structure

As long as you are talking about "which one is the subject," agreement is far away.
What is truly important is "what the reader receives."

What is needed in public relations is not the technique of paraphrasing, but a perspective that shifts the starting point of structure and language.
Generative AI helps with that perspective shift.

■ What to do in these situations?

Q: A partner company is aggressively pushing for a "company-centric" draft
→ Use Claude to analyze the draft, use Genspark to visualize "how it will be read from the reader's perspective," and use that as material for discussion.

Q: We are having trouble with the order of company names despite it being a "joint announcement"
→ Ask ChatGPT to generate a draft that presents both company names in a parallel and flat manner.

Q: The tone is inconsistent and it won't come together
→ Use Claude to integrate and analyze past documents, and create a writing style template to unify them.

■ Reader Workshop: Let's create a "Drafting Template for Partners"

1. Organize the characteristics of your company and your partner, and express them structurally using Genspark

2. Generate a draft with ChatGPT that uses "the value received by the customer" as the subject

3. Create document visuals with unified colors for both companies using Canva

4. Check the integration of the writing tone and the balance of expressions using Claude

■ Conclusion

In collaboration, "how it is talked about" is more important than "making an announcement."
Generative AI has the power not only to balance things, but to find a third way of speaking = a subject from the reader's perspective.

What public relations should aim for is not "which one is the subject," but
to keep asking the question, "how will this reach the reader?"

41 | The survey release isn't hitting home... How to redesign it into a structure that conveys the message

(Chapter 5 | Turning diverse outputs into "meaningful communication")

■ Introduction: "The data is interesting, so why isn't it getting through?"

The HR startup "Lifehack Career" conducted an independent survey on the career consciousness of working people in their 20s and 30s.
Expectations were high within the company that it would "become a hot topic as a PR initiative."

Public relations representative Akari Kido created and distributed a survey release, but the results were not favorable.

• Zero inquiries from the media

• Weak reaction on social media

• Getting grilled in internal reviews with, "So, what's your point?"

■ Case: The 'structure' of the argument and the data were misaligned

Kido first uploaded past surveys, competitor cases, and news articles to NotebookLM to analyze the tone and structure. He verified it using the following prompt.

Prompt: "Based on the following past survey articles and news examples, analyze the common structural patterns and how they draw in the reader. Point out the elements missing from our company's release structure."

→ It was discovered that the introduction was weak in explaining "what this data reveals" and was not designed to answer the "reader's question."

Next, he had ChatGPT generate a "structure that moves the reader's emotions" based on the characteristics of the data.

Prompt: "Based on the results of this survey data, redesign it into a structure that 'piques the reader's interest and gets shared.' Organize it in the flow of: Reader's Question → Data Presentation → Interpretation/Implication → Call to Action."

Also, he revamped the diagrams and graphs in Canva, restructuring them so they are "visually communicative" in social media and documents.

■ Solution: Changing the 'reader design' of the survey changes how it is received

In the restructured release, he placed the question "What is the 'moment when young professionals think about changing jobs'?" at the beginning, followed by a flow of: unfolding the survey data → insights → raising issues → corporate perspective.

As a result, there were 8 media placements, and many reactions such as "I relate" and "I get it" were received on social media.

What was important was the design that worked backward from "what I want the reader to feel" rather than "what we investigated."

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt)
NotebookLM | Structural analysis of past survey articles and news cases | "Extract common 'patterns of how they are read' from the following articles"
ChatGPT | Redesigning the release structure to be read | "Restructure the data into a 'structure that reaches the emotions.' Follow the flow of: Interest → Presentation → Interpretation → Action"
Canva | Data visualization and visual revamp of graphs | "Propose graphs and diagrams that look great on social media and in documents, tailored to the content of this release"

■ Column: A survey is a story of 'questions and implications'

A survey release is not a "place to report data," but a "conduit to raise questions and deepen thinking."

Instead of being buried in numbers, by converting them into a story of Question → Hypothesis → Data → Action, the reader changes from a "thinking reader" to an "acting reader."

Generative AI becomes the ideal partner for organizing and redesigning that structure.

■ What should I do in cases like this?

Q: I am told that my conclusions are blurry even though I have the data.
→ Ask ChatGPT: "Suggest multiple conclusions or questions that this data can best appeal to."

Q: My research angles tend to overlap with others and get buried.
→ Use Perplexity to collect "angles on similar themes reported recently" and find points of differentiation.

Q: The graphs are complex and hard to understand.
→ Choose a graph template in Canva, and ask ChatGPT to summarize it by saying, "If you had to narrow this down to three elements, what would they be?"

■ Reader Workshop: Let's create a "Research Release Redesign Sheet"

1. Upload past research releases and competitor articles to NotebookLM to analyze their structure.

2. Use ChatGPT to generate "the most compelling question for the reader -> structural proposal."

3. Revamp diagrams and graphs in Canva, and adjust them for materials and social media.

4. Share internally and refine it into a "structure that prompts action" from the perspectives of multiple departments.

■ Conclusion

A research release is not a list of facts, but a "catalyst for emotion."
If you can restructure the questions and insights behind the numbers from the reader's perspective,
information becomes "words that reach them."

Generative AI is a partner in the exploration of finding the questions and structures that need to be conveyed.

42 | Interview articles are too long... How to find "words that convey everything in one sentence"

(Chapter 5 | Turning diverse outputs into "meaningful communication")

■ Introduction: "It's a good story, but it's long"

Shota Miura,Shota Miurawho works in the public relations department of a major telecommunications company, regularly produced interview articles introducing internal initiatives.

This time, the content covered the challenges of a young engineer. It had passion, and the words spoken were persuasive.
However, when he submitted the completed manuscript to his boss—

Miura slumped his shoulders in disappointment.
Even though he thought he had confidence in his "ability to convey," his own words at that moment lacked the "ability to make people read."

■ Case: "Extracting" instead of cutting

Miura first hadChatGPTread the entire text and extract "a sentence that leaves an impression on the reader."

Prompt: "From the following interview transcript, please propose 10 patterns of 'catchphrases that stay in the reader's heart.' Please preserve the speaker's tone without compromising the meaning."

→ Among the proposed sentences, the very words that were impressive during the interview were sharply condensed.

Furthermore, Claude was utilized to redesign the structure of the entire article to make it easier to read.

Prompt: "Please restructure this interview article into a 'composition that makes the reader feel empathy and conviction.' To ensure readers don't ask 'So what?', please clearly state the purpose of each paragraph."

With this approach, I was able to refine the article into a structure that 'gets through' rather than just 'conveys'.

■ Solution: Think about 'what to keep' rather than 'what to cut'

The biggest lesson Miura learned was not to 'cut words' but to 'bring the core message to the surface'.

• An interview is a 'stratum of speech.' By digging, the core becomes visible.

• AI acts as a 'third-party reader perspective,' discerning the nuances of the words.

• An impressive catchphrase can become the 'backbone' of the entire structure.

As a result, the article received praise from my boss, saying 'The title made me want to read it' and 'It leaves a good impression after reading,' and it became an article that gathered many comments when shared internally.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Used Prompt (Excerpt) | ChatGPT | Extracting catchphrases | "Propose 10 catchphrases from the full text that 'stay in the reader's heart' while maintaining the speaker's tone" | Claude | Reviewing structure and clarifying purpose | "Create a structure so that readers don't think 'So what?'. Please give each paragraph a 'role'"

■ Column: Editing techniques to measure the 'density' of words

The reason a manuscript becomes long is not because 'there is too much to convey.' It is because
you have not identified which words are the core.

Generative AI is also a 'presence that compares the weight of words on behalf of people.' When in doubt, ask the AI: 'Which words should remain?'

Not 'cutting,' but 'keeping.' That is the essence of editing in the AI era.

■ What to do in these situations?

Q: I feel bad about cutting the other person's words, so the text becomes long.
→ Try giving ChatGPT the instruction: 'Summarize it to be short and powerful while maintaining the speaker's own voice.'

Q: I can't tell for myself which parts of the manuscript are 'unnecessary.'
→ Please ask Claude: 'Looking at this from the reader's perspective, please point out unnecessary paragraphs or overlapping content.'

Q: The flow of my interviews is inconsistent every time.
→ By setting up a "question template" in Notion and using ChatGPT to articulate the intent behind that order, you can keep your focus steady.

■ Reader Workshop: Let's practice "conveying in a single sentence"

1. Have ChatGPT read your past company interview articles and extract a "sentence that resonates with readers."

2. Check if the flow of the entire article feels cohesive when you use that sentence as the title.

3. Organize the structure with Claude, articulate the purpose of each paragraph, and read it through again.

4. Adjust any paragraphs that lack impact or parts that are repetitive.

■ Conclusion

What is needed for an interview article is not "telling everything."
What to leave unsaid, and where to fully convey your message.

Generative AI is surprisingly helpful for the "reader-perspective editing" required for that.
May the words you write become a "single sentence" that remains in someone's heart.

43 | Podcast Delivery Scenarios: How to Use AI to Polish Your Content for Listeners

(Chapter 5 | Turning Diverse Outputs into "Meaningful Communication")

■ Introduction: "The content is interesting, so why isn't it being listened to?"

RelayLab, an education-focused startup in Tokyo. PR representative Mari Okano started a podcast program to share the company's worldview and work style.

The theme is "Education x Technology." They invite the CEO and employees as guests and release a 20-minute interview-style episode every month.

She was confident in the content, but the number of plays remained stagnant.

• "The content is good, but people say the intro is too long."

• "It's hard to tell where the highlights are."

• "There are many comments asking, 'What are you trying to say in the end?'"

Okano realized, "Podcasts are 'content read with the ears.' They require a different kind of structural skill than writing."

■ Case: Moving toward a structure that draws an "auditory map" rather than just text

Okano first had ChatGPT read the transcripts of past episodes and entered the following prompt.

Prompt: "Please redesign the following podcast transcript into a structure that ensures listeners do not get lost. Evaluate whether there is a hook at the beginning, if the flow of the conversation has ups and downs, and if it leaves a lasting impression at the end."

→ It was pointed out that there was "ambiguity in the introduction" at the start, "tangents" in the middle, and "weak closing" at the end.

Next, he customized the structure template using Claude.

Prompt: "For a podcast themed on education x technology, please propose the most suitable structural elements for the 'first 10 seconds, middle, and end,' based on past structures."

Furthermore, he organized a 'scenario question template' to use before recording with Notion AI. With Canva, he also revamped the thumbnail images.

■ Solution: 'Subtractive design' for the ears creates a structure that gets listened to

What Okano learned was the 'order of conveying information' and the 'framework that supports the tempo'.

• 'Vocal inflection' is more important than the text itself

• Include a line at the beginning that conveys the 'value of listening now'

• Design the end with the intention of leaving 'words that stick'

As a result, the reconstructed episode saw double the number of plays. Reactions on social media also increased, with comments like 'easy to listen to' and 'I was able to listen to the end'.

■ Summary of AI tools used and prompts

Tool Usage Scenario Used Prompt (Excerpt) ChatGPT Analysis of structural issues "Read the podcast transcript and change it into a structure that is easy for listeners to understand" Claude Organizing structural templates for the beginning, middle, and end "Please propose an optimal structure template for an education x tech podcast" Notion AI Preparing scenario question templates "Please organize a list of questions for podcast recording in the order of introduction -> deep dive -> summary" Canva Revamping thumbnails for distribution "Please create 3 thumbnail proposals that convey the content of the podcast in one phrase"

■ Column: The ears listen by 'predicting'

People try to 'read ahead' and understand audio information more than written text.

That is precisely why, in a podcast, you need to convey 'What is this about?' and 'What will I gain?' at the very beginning.

Generative AI is a presence that can refine whether there is waste in the structure and whether the flow of the conversation is natural, from the 'listener's perspective' rather than the reader's.

'Audio is a graph of emotions'—it is the role of PR to support that graph with the line of structure.

■ What to do in these situations?

Q: My conversations always go off-track.
→ Prepare a "question template" in Notion and use Claude to create a structure that allows you to "dig deep while still being able to return to the main topic."

Q: I feel like no one in the company listens even after I publish.
→ By creating a title image in Canva that "lets people imagine the content," you can increase visibility on internal social media and other platforms.

Q: I'm told my opening is weak.
→ Have ChatGPT read your transcript and suggest multiple introductory hooks that "make people want to listen within 10 seconds."

■ Reader Workshop: Let's organize your company's podcast structure

1. Have ChatGPT read past transcripts to check for structural improvements

2. Have Claude suggest the optimal structure template for the "beginning, middle, and end"

3. Prepare "pre-interview questions" and "flow of conversation" templates in Notion

4. Create multiple "eye-catching thumbnails" in Canva and have them voted on and improved within the company

■ Conclusion

Podcasting is "spoken" PR. That is precisely why both the weight and lightness of words are conveyed.

Organizing the structure is nothing less than "closing the distance between the speaker and the listener."

Generative AI, as an editor with a listener's perspective, will gently support your "power to tell a story."

44 | Designing YouTube video descriptions to avoid being "skimmed over"

(Chapter 5 | Turning diverse output into "meaningful communication")

■ Prologue: "I thought the description was being read properly"

HomeLinks, a renovation-related SaaS company. Shota Ishida from the marketing team was in charge of managing the YouTube channel in collaboration with the PR team.Shota Ishida was in charge of managing the YouTube channel in collaboration with the PR team.

The company was focusing on video production, regularly distributing construction case studies, customer interviews, and product introductions, but—

• "The view count is decent. But it doesn't lead to inquiries."

• "There are few comments. It's not being shared either."

• "Is it possible that no one is actually reading the video's 'description'?"

Meanwhile, a series of viewer surveys revealed that many people "did not notice the links in the description box."

"The video content is good, but it's a waste,"
Ishida felt, and decided to rethink the design of the description text itself.

■ Case: The description text is "another video content"

First, Ishida had ChatGPT read the description boxes of 10 past YouTube videos and fed it the following prompt.

Prompt: "Please redesign the following YouTube description text into a structure that makes viewers want to take the next action. Please suggest what kind of flow should be created according to the three timings: before, during, and after viewing."

→ It was pointed out that many description boxes ended with "just a description of the content" and that the call to action (CTA) was weak.

Next, he had Claude organize the ideal structure for a description text as follows.

- Beginning: Present the main points of the video and the benefits to the viewer in one sentence - Middle: Detailed content, chapter breakdown, and summary of points - End: CTA (inquiries, document downloads, links to other videos)

Furthermore, he used Notion AI to organize topics and extract keywords for each video, and used Perplexity to analyze the description text structures of popular channels from other companies.

■ Solution: Doubling the action rate with a description text that "makes people want to read"

The redesigned description text was structured not just as an explanation, but as "another video."

• Include a "valuable message" and the significance of watching in the first three lines

• Organize "highlights" and "chapter breakdowns by time" in the middle

• Always clearly state a "CTA to encourage action" at the end

As a result, document requests from the description box doubled, and the retention rate after viewing also improved.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt) | ChatGPT | Redesigning description structure | "Please organize the description text into a structure that makes viewers want to take the next action." | Claude | Templating the ideal structure | "Please organize the necessary components for a YouTube description text for before, during, and after viewing." | Notion AI | Summarizing video content and extracting keywords | "Based on this video content, please extract the important keywords and points that should be included in the description text." | Perplexity | Researching structures of other companies' cases | "Tell me the trends and best practices for description texts used on YouTube channels of SaaS companies."

■ Column: The description box is the "last mile of text"

Viewers actually look at the description box more than you might think. However, they judge whether it is "worth reading" in a few seconds.

The design of the description text is truly the "last mile." It is the final flow that determines whether or not to "connect" the video you worked so hard to deliver to inquiries or document downloads.

Generative AI acts as a partner that helps polish your writing by visualizing 'what gets read' and 'where readers drop off' from the reader's perspective.

■ What to do in these situations?

Q: The summary is too long and doesn't get read
→ Use Claude to create a '3-line summary' and place it at the beginning.

Q: People say they don't know the highlights of the video
→ Organize the chapters in Notion and use ChatGPT to generate a 'timeline with highlights'.

Q: The call to action (CTA) is weak
→ Have ChatGPT generate multiple CTA options and use Perplexity to design them by referencing competitor examples.

■ Reader Workshop: '1 Video x 1 Summary' Practice Book

1. Select one of your company's YouTube videos and use Notion to transcribe and organize the content

2. Use Claude to create a template for a '3-line summary + call to action' structure

3. Have ChatGPT generate multiple summary drafts and review them internally

4. Use Perplexity to compare your structure with other companies' channels and generate improvement ideas

■ Conclusion

No matter how good a video is, it's meaningless if it isn't watched.
And if it doesn't lead to action, you can't say you've 'fully delivered' it.

The description section is both a bridge for information and an entrance for action.

Together with generative AI, move from 'showing' to 'driving action'.
That is the next step required of public relations.

45 | Update the IR page! Maintenance techniques that balance readability and reliability

(Chapter 5 | Transforming diverse outputs into 'meaningful communication')

■ Introduction: 'It exists, but I don't want to read it'

Daiei Tech, a listed manufacturing company.
Ryo MiuraRyo Miurawho is also in charge of managing IR content in the PR department, was entrusted with the role of liaison with corporate planning.

On the IR page,

• PDFs of financial results and explanatory materials

• Shareholder meeting-related information

• Integrated reports and ESG data

are lined up matter-of-factly. The information is all there. But—

It’s “there,” but people “don’t want to look at it.”
How to change such an IR page—Miura began utilizing AI while struggling with this dilemma.

■ Case: IR information only “conveys” when it is “edited”

Miura first had NotebookLM read through three years of past IR-related PDF documents and the proposed page structure.

Prompt: “Read the following IR-related materials and web page structure, check if it is in a ‘structure that is easy to understand’ for investors and stakeholders, and provide improvement suggestions.”

→ The points pointed out were as follows:

• Information is ‘stored,’ but there is no ‘navigation path’

• There is no consistency in moving between PDF and HTML, leading to many exit points

• The ‘company’s will and stance’ are reflected in the ‘order of items’ rather than in words

So, Miura asked ChatGPT to redesign it as a ‘sample structure for when the integrated report is converted to HTML.’

Also, Canva was used to generate KPI graphs and visuals for IR topics.
Copilot was used to create prototype plans for website renovation, which were utilized in internal sharing materials.

■ Solution: The IR page is not a ‘financial portal’ but the ‘face of the company’

What Miura realized was that the IR page needs the ‘power to attract people.’

• What are investors coming to look for?

• What kind of design makes people want to look at it repeatedly?

• Where does ‘trust’ come from?

He verbalized these with AI and created an ‘IR Design Policy Document.’
While building consensus within the company, he started the restructuring project.

■ Summary of AI Tools Used and Prompts

Tool | Use Case | Prompt (Excerpt)
NotebookLM | Structural analysis of current IR composition | "Identify the 'drop-off points' and provide 'improvement suggestions' for the structure of this IR page."
ChatGPT | Composition design for HTML version of integrated report | "Convert the IR materials into an HTML structure and propose a 'sequence that is easy to read'."
Canva | Visualization of IR data | "Turn these KPIs and trends into a graph that is easy for investors to understand."

Copilot | Prototyping for IR page renovation | "Based on this structure, create a wireframe proposal for the IR page layout."

■ Column: The IR Page is a 'UI of Trust'

IR is not just 'public information'.
It is, in itself, a 'design of trust' aimed at investors.

Who is coming, and what are they looking for?
If you cannot trace that gaze, it becomes nothing more than a 'file storage'.

UI (User Interface) is the 'grammar of trust'.
Refining that text is a collaborative effort between PR and AI.

■ What to do in these situations?

Q: There are too many IR materials and I can't keep up with organizing them.
→ Create a 'classification map' with NotebookLM and organize the navigation flow with ChatGPT.

Q: Management says 'it's dangerous to cut information'.
→ Use Claude to generate a 'proposal to reorganize by changing the order without cutting information' and use it for explanatory materials.

Q: The structure of the English IR page is inconsistent with the Japanese one.
→ Use Copilot to create a template for the multilingual UI structure and use ChatGPT to align the phrasing.

■ Reader Workshop: Let's create an 'IR Page Redesign Checklist'

1. Upload a list of past IR content to NotebookLM for analysis

2. Have ChatGPT generate multiple 'IR site structure template proposals'

3. Visualize KPIs with Canva and summarize renovation images with Copilot

4. Organize and explain the 'investor behavior flow map' for internal meetings

■ Conclusion

IR is indeed a disclosure obligation.
But that page is also a place that speaks of 'trust' while wearing the face of an obligation.

Readability is sincerity, and flow design is a company's stance.

Polishing that structure with AI is not just a simple web renovation.
It is the 'refining of words' by a company for its future dialogue partners.

46 | FAQ and Customer Support Messages: Refining "Words That Don't Mislead" with ChatGPT

(Chapter 5 | Turning Diverse Outputs into "Meaningful Communication")

■ Introduction: "You've been asked that before, haven't you?"

EC startup "GloNet." Public relations representative
Rio Mochizuki, who also handles user support, was struggling with how to streamline customer service.

• It's a common question, yet the FAQ is hard to find

• The chatbot's responses are off-base

• The support tone is "too stiff," sometimes causing even more anger...

Expressions and templates used repeatedly within the company can sometimes appear to have "different meanings" to the reader. In today's information-overloaded world, isn't it more important to be "not misunderstood" than just to "convey"? Thinking this, Mochizuki set out to improve support using ChatGPT as the core.
Thinking this, Mochizuki set out to improve support using ChatGPT as the core.

■ Case: If the "Structure" is Right, FAQs and Emails Are Less Likely to Be Misunderstood

First, she uploaded existing FAQs, user inquiry history, and template responses to NotebookLM. She identified patterns in common inquiries and areas where misunderstandings easily occur.

Prompt: "Based on the following FAQs and customer support history, identify expressions, structures, and sequences that are prone to misunderstanding. From the 'reader's perspective,' provide concrete examples of expressions that should be improved."

Next, she restructured the text for the 10 most frequently asked questions using ChatGPT.

Prompt: "Restructure the following content so that the 'reader will not misunderstand.' Please use simple language, short sentences, and a logical sequence, ensuring the main points are conveyed even if the reader skims the text."

Furthermore, she improved FAQ tag classification and searchability with Notion AI. She created Q&A slides with images using
Canva, organizing them into a format that can also be used for internal training.

■ Solution: It Was "Structure," Not "Vocabulary," That Prevents Misunderstanding

Mochizuki realized that no matter how carefully you write, "if the structure is bad, it won't get through."

• Just changing the order of sentences reduces frustration

• If the "premise" of the answer is visible, it provides peace of mind

• Paying attention to "connective words" rather than sentence endings reduces misunderstandings

Based on these insights, I generated a "User Response Rewrite Template" using Genspark. It was established as a "common language" between the PR team and the CS team.

■ Summary of AI Tools Used and Prompts

Tool | Use Case | Prompt (Excerpt)
NotebookLM | Extracting misunderstanding risks from FAQs and history | "Extract expressions and structures that are easily misunderstood and provide improvement suggestions."
ChatGPT | Reconstructing text from the reader's perspective | "Reconstruct the FAQ so that the reader 'does not misunderstand'."
Notion AI | Improving FAQ categorization and search navigation | "Categorize the FAQ with tags and create a structure that is easy to search from a user's perspective."
Canva | Creating Q&A slides with images | "Create Q&A slides that visualize the CS text."
Genspark | Creating text templates | "Output a 'collection of templates for writing that is hard to misunderstand' for FAQ and CS text."

■ Column: Customer communication is not about 'apologizing,' but about 'providing reassurance'

What is important in customer support is not the cliché of 'We apologize for the inconvenience...'.
It is whether the structure allows the reader to 'feel at ease'.

• The location of the problem is immediately visible

• The response procedure can be read immediately

• Alternatives and timelines are provided

Does this 'proactive reassurance' permeate the words?
That is the first step in preventing misunderstandings.

■ What should I do in these situations?

Q: I am told that my responses are just templates and lack 'warmth'.
→ Use ChatGPT to generate a 'version with a human touch' and adjust the tone while comparing them.

Q: There are many pages in the FAQ that are not being accessed.
→ Use NotebookLM to analyze the 'reasons for lack of access' from both the content and navigation perspectives.

Q: The language used by CS and PR is inconsistent.
→ Use Genspark to create a 'Response Tone Unification Template' and promote the use of a common language.

■ Reader Workshop: Create a 'Misunderstanding-Free Text Template' for your company

1. Load existing FAQs and inquiry history into NotebookLM to visualize trends

2. Generate multiple samples of 'text that does not cause misunderstandings' using ChatGPT

3. Output 'recurrence-prevention style expression templates' using Genspark

4. Organize them into Q&A cards for CS team training using Canva

■ Conclusion

FAQs and customer response texts are also the front line of public relations.
For users, that text can sometimes become the 'personality of the company'.

Not just "being read," but "being understood" and "being trusted."
The power of AI is indispensable for designing such language.

Reducing misunderstandings is about building up trust.
By using AI to refine the "structure of communication," words become much more effective at reaching people.

47 | How do you translate cultural nuances in English documents for overseas distribution?

(Chapter 5 | Transforming diverse output into "meaningful communication")

■ Introduction: "The grammar is correct. But something feels off."

Medical SaaS company "MediLink."
With expansion into Southeast Asia just around the corner, the person entrusted with global PR wasYosuke Yamashita. He is good at English, but...

• Pointed out as being "mechanical and cold" due to a direct translation style

• Told by a local PR firm to "be more culturally sensitive"

• Vague feedback from native English speakers like, "I understand the meaning, but... it doesn't quite click."

Differences in cultural "assumptions" change the intent of the words.
Yamashita begins redesigning language and culture using AI, not for "translation," but for "effective localization."

■ Case: Visualize even the "context of meaning" that cannot be conveyed by translation alone

First, he compared and analyzed past press releases translated between Japanese and English usingChatGPT.

Prompt: "Compare the following Japanese release and its English translation, point out 'areas where the intent may not be correctly conveyed,' and suggest improvements. Please also add comments from a cultural and expressive perspective."

Next, he organized the "differences in how things are conveyed" from the perspective of multilingual culture usingClaude.
For example, it was pointed out that Japanese honorific expressions like 'okagesama de' (thanks to you) could potentially 'look dependent' in English.

Furthermore, he usedPerplexityto investigate the expression trends in overseas materials and IR releases of competitors.
He identified subtle differences across cultural spheres, such as "phrasing of gratitude" and "humble ways of conveying achievements."

Finally, he generated a "culture-specific English translation prompt template" usingGenspark.
He made it so it could also be used for communication with local PR firms.

■ Solution: From "translation" to "communication design"

The biggest realization Yamashita gained was the importance of "conveying the assumptions" rather than "translating the words."

• Have AI visualize the "context" rather than just the words

• Honorifics, consideration, and modest expressions require a 'redesign of meaning'

• Position English translation as a 'starting point,' not a 'goal'

By understanding how things are 'read' in each culture and adjusting language through dialogue with AI, it has become possible to 'communicate in a way that resonates with the other party.'

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Prompt (Excerpt)
ChatGPT | Comparing and adjusting the intent of Japanese-English translations | 'Compare this Japanese and English text, and point out and improve any nuances that are not being conveyed.'
Claude | Analyzing expressions from a multicultural perspective | 'Explain how this expression would be perceived by a native English speaker from a cultural perspective.'
Perplexity | Researching overseas releases and materials from other companies | 'Tell me about the trends in phrasing used in overseas PR materials by medical SaaS companies.'
Genspark | Translation and localization templates by cultural sphere | 'Output culturally sensitive templates that can be used for Japanese-to-English localization, categorized by country.'

■ Column: Localization means respecting the 'thought structure' of the other party

Translation is the technology of carrying 'meaning.'
Localization is the technology of delivering 'premises' along with it.

For example, the Japanese 'moushiwake gozaimasen' (I am sorry) can look like an 'admission to avoid litigation' in the English-speaking world.
If the background of an 'apology' is different, the 'reception' is also completely different.

What is important is to imagine 'in what context people of that culture will receive those words.'
Generative AI is a powerful assistant that can give shape to that 'imagination.'

■ What should I do in these situations?

Q: I am always told that my writing style is too stiff in overseas releases
→ Use Claude to polish it into 'natural native English' and use ChatGPT to compare the tone.

Q: We are in the early stages of global expansion and have no in-house expertise in localization
→ Use Perplexity to collect examples and trends from other companies in the industry, and use Genspark to create templates.

Q: I am told that the meaning is 'off' every time I finish with just automatic translation
→ It is effective to proceed in a dialogue format with ChatGPT, confirming step-by-step from intent to expression to syntax.

■ Reader Workshop: Creating 'Effective Expression Templates by Cultural Sphere'

1. Verify and compare existing Japanese-to-English translations using ChatGPT

2. Identify areas for improvement from the perspective of 'cultural nuance' using Claude

3. Collect global PR expressions from other companies using Perplexity

4. Generate a 'Cultural Template Collection' using Genspark and deploy it within your company

■ Conclusion

When the person you are communicating with changes, the words change too.
The role of global PR is to refine those changes while respecting them as 'cultural differences.'

Translation can be done by AI.
However, designing "words that resonate" is born from the dialogue between humans and AI.

Approach culture with respect, and use imagination to go beyond mere translation.
That is the true thrill of public relations, which "connects the world through words."

48 | How to express "consistent messaging" for external roles and concurrent positions?

(Chapter 5 | Transforming diverse outputs into "meaningful communication")

■ Introduction: "Wait, who is this person?"

Healthcare startup "Well-B Inc."
PR representative Jun Kasahara felt uneasy one day after seeing an inquiry from outside the company regarding a featured article.

In the company's role introduction, he was listed as "CTO"; on his social media profile, as "Engineer, Researcher, Entrepreneur";
and at speaking events, he was introduced as "Social Entrepreneur / Well-B CTO."

Kasahara realized that his "persona" was being conveyed inconsistently across media and internal/external communications.
He was struggling with "how to communicate when there are multiple roles."

■ Case: Organizing with a "story" rather than job titles

Kasahara first uploaded his past speaking engagements, contributions, press releases, and social media posts to NotebookLM.
He explored whether there was consistency in his "roles" and "philosophy" based on his words as CTO and his activity history.

Prompt: "Analyze the following text and organize 'what beliefs and vision this person acts upon.' Also, propose a title and introduction template to ensure consistency in internal and external communications."

Furthermore, he input the following prompt into ChatGPT:

Prompt: "For a person with the following multiple titles (CTO / University Lecturer / Foundation Director, etc.), reconstruct an introduction as a 'personal story unified by a single axis.' Please use expressions that allow readers to understand 'who this person is' at a glance."

By weaving it back together as "one story" rather than a list of titles, a sense of unity was born in the message.

■ Solution: "Who is this?" should be told through "interests," not "titles"

What Kasahara realized was that "personal introductions should be conveyed through 'will,' not specs."

• Summarize multiple activities by "values" rather than "job functions"

• Maintain consistency between the tone of the communication and the introduction

• Design based on "how one will be remembered" when viewed from the outside

Based on this design, I generated a "Multi-role Introduction Template" using Genspark. After that, I standardized the introduction text for all executives within the company and shared it as a guideline for various media outlets.

■ Summary of AI Tools and Prompts Used

Tool | Use Case | Prompt (Excerpt) | NotebookLM | Organizing and analyzing the philosophy behind speaking engagements, articles, and social media posts | "Organize the consistent philosophy and values of this person and apply them to the introduction text." | ChatGPT | Reconstructing introduction text | "Reconstruct this person with multiple titles as a story based on a common purpose." | Genspark | Generating introduction templates | "Create a multi-role personal introduction template. Use a structure that maintains consistency across titles."

■ Column: An introduction is not a "translation of titles" but a "summary of values"

When introducing someone, simply listing their titles does not convey the "meaning."

For example, the titles "University Lecturer, Technical Advisor, and Startup CTO" might be summarized as a "practitioner interested in the future of technology and people."

An introduction is a piece of text that creates an image in the reader's mind of "this person might be like this."

Together with AI, scoop up the "will" behind the titles and shape it into something that resonates. That is the "personal design capability" required of modern PR.

■ What to do in these situations?

Q: I have many concurrent roles, so I can't settle on a self-introduction.
→ Use ChatGPT to refine it into an "introduction text centered on common thoughts and values."

Q: Multiple profiles exist in a scattered way, and I can't manage them.
→ Consolidate past communications in NotebookLM to unify the tone and context.

Q: I want to make it easier to reuse introduction text both inside and outside the company.
→ It is convenient to create a template in Genspark and manage it in a format that is easy to re-edit.

■ Reader Workshop: "Refining an introduction that conveys the 'multi-faceted talent' of your company"

1. Upload past articles, speaking engagements, and profiles to NotebookLM.

2. Create an introduction text based on common intentions or themes using ChatGPT.

3. Turn it into a reusable "introduction template" in Genspark.

4. Share it within the company so that consistency of the "personal image" can be maintained when communicating externally.

■ Finally

PR is also the work of conveying to society that "this company has these kinds of people."

The more titles you have, the harder it becomes to convey who you are. But behind those increased titles, there is always "one single thought."

By weaving an introduction centered on that passion, you create PR that makes people think, 'I want to meet this person.'

49 | Shareholder Newsletters: Techniques for Reconstructing from an 'Investor Perspective'

(Chapter 5 | Transforming Diverse Outputs into 'Meaningful Communication')

■ Introduction: Who listens to a 'newsletter' filled only with numbers?

Technicolab, a B2B subsidiary of a long-established manufacturing company, was struggling.
Tsubasa Hasegawa, who handles both PR and IR, Tsubasa Hasegawa was having a hard time creating the quarterly 'Shareholder Newsletter'.

• We publish it every quarter, but there is no response.

• No feedback arrives, so there is no way to improve it.

• It is unclear 'what kind of information should be included'.

Past newsletters were mainly lists of facts, such as financial summaries, numerical trends, and lists of new products.
However, it had become unclear 'who it was being written for'.

■ Case: Reordering in the 'Order the Reader Wants to Know'

Hasegawa first used Perplexity to collect examples of shareholder communications and newsletter structures from other companies.
He particularly referred to IR materials from startups and companies about to go public.

Next, he uploaded his company's past newsletters to NotebookLM to analyze the tone and structure, utilizing prompts like the following:

Prompt: 'Analyze the following shareholder newsletter and extract the missing elements and strengths of what investors are interested in. Also, present an ideal structure proposal reordered in the sequence that interests the reader.'

Furthermore, he asked ChatGPT to reconstruct the text with the following prompt:

Prompt: 'Based on the following content, please reconstruct it into a 'narrative style' that is easy for shareholders to read and makes them look forward to the future. Replace difficult terms and figures with easy-to-understand ones, and end with a closing that encourages the next action.'

■ Solution: Toward Communication with a 'Narrative' that Conveys Future and Emotion

What Hasegawa arrived at was the perspective that 'Shareholder communication is a letter conveying where the company is heading'.

• Talk about numbers as 'proof of the future'

• Give a sense of story to 'why we are making this choice now'

• Speak in 'human language' like a single letter

Based on this concept, I used Genspark to create an 'Investor Story Structure Template' to serve as the foundation for future communications.

■ Summary of AI Tools and Prompts Used

Tool | Use Case | Prompt (Excerpt) | Perplexity | Collecting and analyzing IR materials and shareholder newsletters from other companies | 'Please provide examples of shareholder newsletter structures that investors prefer. Especially examples from startups and listed companies.' | NotebookLM | Analyzing our own newsletter | 'Reorder this in a way that interests the reader, and visualize the strengths and weaknesses.' | ChatGPT | Converting to an easy-to-read tone | 'Rewrite this in a tone that resonates with investors and makes the text build anticipation for the future.' | Genspark | Templating and designing future operations | 'Please create a "narrative structure template" for shareholder newsletters.'

■ Column: An Era Where IR Also Requires a "Narrative"

What a company needs to tell is not just "what we did," but "why we did it."

Numbers and product information can be found by searching.
However, the "reasons for decisions" and "future intentions" behind them are rarely told in IR materials.

That is why it is important for newsletters to have a narrative.
A story that makes people think, "I want to support this company," turns shareholders into "fans."

■ What should I do in these situations?

Q: Our business is niche and I can't explain it well.
→ If you tell ChatGPT to "explain it simply for the general public," it will express it using metaphors and examples.

Q: The roles of IR materials and shareholder newsletters overlap.
→ Use NotebookLM to analyze past materials for both and clarify the overlaps and complementary relationships.

Q: I want to add a "narrative" to a newsletter full of numbers, but it's difficult.
→ It is convenient to create a "qualitative + quantitative story structure template" using Genspark.

■ Reader Workshop: Reconstruct "Your Company's Shareholder Newsletter" with a Narrative

1. Upload past newsletters to NotebookLM and organize the trends in structure and expression.

2. Collect examples of IR stories from other companies using Perplexity.

3. Convert your company's news into a "tone that reaches people" using ChatGPT.

4. Create a template with Genspark and establish it as an internal guideline.

■ Conclusion

A shareholder newsletter is like a love letter addressed to "partners walking into the future together."

That is why it needs a story and warmth, not just a list of facts.
Generative AI has the power to turn the intentions and emotions behind the numbers into "words that resonate."

Shape your company's passion into something that truly reaches people.
That is your job at the intersection of PR and IR.

50 | Collaborative PR with Local Governments: How to Bridge the Language Gap Between Public and Private Sectors?

(Chapter 5 | Transforming Diverse Outputs into 'Meaningful Communication')

■ Introduction: Between Private Sector Speed and Administrative Caution

Healthcare venture 'SynMedica'. A health promotion project with a local government had begun, and
Takuto Mochizuki, who is in charge of PR, was starting preparations for the partnership announcement.

However, he struggled more than he expected.

• The administrative approval process is complex

• The text is in 'bureaucratic language' and hard to understand

• It lacks the warmth of a private company

'Even though it's an announcement that we are working together, it doesn't feel exciting when you read it.' The entire project team shared this sense of discomfort.

■ Case: Visualizing Warmth and Intent Through 'Language Translation'

Mochizuki first fed the draft provided by the local government and past public-private partnership examples into NotebookLM.

Prompt: 'Please convert the following local government announcement into text that is easy to understand and approachable for private sector readers. Adjust the tone so that both the public stance and the intent to collaborate coexist.'

Next, he input a prompt into Claude to check the impression of the text from the three perspectives of the partner company, the government, and the residents:

Prompt: 'Please tell me what kind of impression you get when reading this document from the following three positions: 1. Local government official, 2. PR representative of a private partner company, 3. Citizen/Resident. Please analyze how easy it is to understand and how it resonates emotionally from each perspective.'

As a final adjustment, he gave the following prompt to ChatGPT and restructured it:

Prompt: 'Please restructure the following draft for a public-private partnership project into an announcement that balances 'public nature' and 'friendliness.' Avoid difficult terms and ensure that consideration for the other party's position is conveyed.'

■ Solution: 'Co-creative PR that Leverages the 'Identity' of Both Sides'

The project text was reborn with a tone that balances 'administrative reliability' and 'corporate speed.'

• Public-like accuracy is guaranteed

• Private-like warmth and story are added

• A context that becomes 'personal' for the public as well

Mochizuki turned this process into a template and shared it internally as a 'Co-creation PR Framework' that could be used for future collaborative projects.

■ Summary of AI Tools and Prompts Used

Tool | Usage Scenario | Prompt Used (Excerpt)
NotebookLM | Reading municipal documents and extracting rewrite material | 'Please translate the municipal announcement into a private-sector PR style. Please also consider the tone.'
Claude | Analyzing post-reading impressions from three perspectives | 'Please analyze the emotional impressions from the three standpoints of government, private sector, and the public.'
ChatGPT | Reconstructing announcements to balance public interest and approachability | 'Please rewrite this public-private partnership press release, balancing 'formality' and 'warmth'.'

■ Column: Adjusting the 'Temperature' of Words with AI

The government and the private sector use different words.
Their goals, organizational cultures, and sense of speed are also different.

However, words have common ground.
If you start from 'who you want to reach and how you want them to feel,' differences can be 'translated.'

Generative AI is that translator.
By adjusting the temperature and intent between the lines through AI, co-creation begins with words.

■ What to do in these situations?

Q: I feel the text created by the government is 'too stiff to use'
→ Use NotebookLM to read the government documents, and use ChatGPT to convert them into 'language that reaches the public'.

Q: I'm told to use government text as-is internally, but it feels wrong...
→ Use Claude to perform a multi-perspective impression analysis and show 'why it doesn't get the message across' using data.

Q: How should I write expressions that reconcile multiple standpoints?
→ Have ChatGPT structure it with a prompt that focuses on 'text that communicates to three parties'.

■ Reader Workshop: Creating a 'Temperature Adjustment' Template for Public-Private Partnership Documents

1. Upload municipal documents to NotebookLM and analyze the structure and tone

2. Rewrite into 'expressions that the public can also read' using ChatGPT

3. Perform an impression analysis from three perspectives using Claude to find areas for improvement

4. If you turn this process into a template, it can be applied to the next project as well

■ Conclusion

Public-private partnership is a collaboration where 'the goal is the same, but the culture is different.'
The key to overcoming those differences lies in 'word design.'

Generative AI supports language that considers the other party's position and the reader's sense of temperature.
It is not just translation, but building a bridge for communication.

To walk in the same direction as both the administration and the public.
Let's start by aligning the warmth of our words.

51 | CSR x Regional Collaboration: PR that conveys a sense of 'co-creation' rather than 'support'

(Chapter 6 | Cultivating Brand Identity through Words and Culture)

■ Introduction: Does your goodwill sound condescending?

Logistics startup 'Tralinks'.
To commemorate its 10th anniversary, the company launched a CSR project to support local children's cafeterias.

PR representative Sayaka Iwanaga was creating an internal newsletter and a press release, but suddenly felt a sense of discomfort.

Even if the thoughts behind the activities are sincere, if you get the 'way of conveying them' wrong, it becomes an imposition.
Can we communicate in a way that doesn't make goodwill a 'one-way street'?
Iwanaga's worries deepened.

■ Case: Designing co-creation from the 'structure'

Iwanaga first fed the following prompt to ChatGPT to check the current text.

Prompt: 'Please change the following CSR announcement text into an expression that conveys a sense of co-creation. Avoid the tone of "the company doing something for the region" and restructure it based on the perspective of "growing together and supporting each other."'

Next, she used Claude to simulate how a third party would feel when reading it.

Prompt: 'Please analyze how one would feel and what impression one would get when reading the following document from the "perspective of the public" and the "perspective of the supported organization." Please comment by focusing on the difference between a "sense of support" and a "sense of co-creation."'

As a result, it was found that words in the text such as "support," "contribution," and "social responsibility" were acting as "words that create distance."

Instead of just cutting words, she used Canva to incorporate photos of the cafeteria site and voices of local children, and rewrote the structure to speak from the "perspective of the field."

■ Solution: Just changing the order of words and the perspective of the subject changes how it is conveyed

What Iwanaga realized is that the difference between "support" and "co-creation" lies in the "order of speaking" and the "position of the subject."

• By placing the voices of the region first, the company becomes an "entity working together"

• The impression changes significantly just by changing the tone from "we did it for you" to "we did it together"

• By clearly indicating "who is the subject" through visuals, warmth is added to the text

Based on this realization, I updated the CSR document template and reflected it in the internal manual.
I have created a foundation that can be applied to all future regional collaboration PR.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Prompt Used (Excerpt)
ChatGPT | Co-creative reconstruction of CSR documents | "Please rewrite this into a CSR release that conveys a sense of co-creation rather than a sense of support."
Claude | Impression analysis from multiple perspectives | "Please analyze the impression of this text when read from the perspective of the beneficiaries and local residents."
Canva | Reflecting field perspectives visually | "Please provide visual ideas that convey the warmth of supporting local children's cafeterias."

■ Column: In CSR, the 'order of the narrative' is everything

No matter how sincere the effort, if you get the 'way of telling it' wrong, your intentions can be conveyed in reverse.
What is important in CSR PR is to stop 'narratives that start from the corporate perspective'.

What should be told first are the regional issues, the voices of those being supported, and the reality on the ground.
The company's thoughts should be told as an 'existence that stands by' those things.

Generative AI can adjust that 'order' and 'choice of words for the perspective'.
Co-creation starts with the structure of the narrative.

■ What should I do in these situations?

Q: I feel like my introduction of CSR activities is coming across as 'condescending'.
→ Use a prompt in ChatGPT to convert the structure from 'supporting' to 'moving forward together'.

Q: The text is well-written, but it feels cold.
→ Performing an 'emotional analysis from a third-party perspective' with Claude will reveal the unconscious tone.

Q: The atmosphere of the field isn't conveyed, and it tends to be symbolic text.
→ By designing field photos + keyword captions in Canva, you can create expressions that carry human warmth.

■ Reader Workshop: Create a 'CSR PR Co-creation Template'

1. Convert your company's CSR documents into a co-creative structure using ChatGPT

2. Confirm the impression from the perspective of beneficiaries and local residents using Claude

3. Supplement the field perspective visually using Canva

4. Share and reuse as a template within the company

■ Finally

CSR PR is not about conveying the 'good things' a company does, but a narrative that stands by the 'reality' of society.
And that narrative asks, 'Where do you start, and whose voice do you start with?'

Generative AI is a partner that questions the structure and the subject of the narrative.
So that your words can reach 'someone's heart'—
Let's evolve from 'support' to 'co-creative' PR.

52 | ESG Reports: From 'unreadable documents' to 'content that gets read'

(Chapter 6 | Cultivating Brand Identity through Words and Culture)

■ Introduction: It was disclosed. But no one read it.

Construction machinery manufacturer "Kousei Engineering".
Public relations representative Ryotaro Otsuka felt relieved after successfully publishing the annual ESG Report.
However, a comment he heard at a meeting a few weeks later clouded his mood.

It didn't become a topic of conversation within the company, and there was little reaction from outside.
The IR department even said, "It was released as a document, but it's not structured to convey the message."
Despite having created it carefully, the emptiness of it "reaching no one"—
Otsuka decided to rethink the way the report should be.

■ Case: Transforming from a "Report" to a "Story"

First, Otsuka had NotebookLM read last year's ESG report and analyzed its structure using the following prompt.

Prompt: "Is this ESG report structured in a way that is meaningful to the reader? Please check for content duplication, information gaps, and the flow of the structure, and suggest improvements to make it a 'document that gets read'."

Next, he redesigned it into a "communicative structure" using ChatGPT.

Prompt: "Please restructure the ESG report below following a story structure. Develop it in the order of 'Challenge -> Initiative -> Realization -> Future', incorporate the voices of employees and the community, and make it easy to read and compelling."

Furthermore, he used Perplexity to compare it with "widely read ESG documents" from other companies.
He extracted trends in tone, use of illustrations, and chapter structures, and applied them to improve the presentation.

■ Solution: The report is reborn as a "Story"

ESG is not a "compilation of data," but a story of "corporate integrity."
By organizing the structure with ChatGPT and NotebookLM and adding human voices to the actual initiatives,
the inorganic report was reborn as "content that makes you want to keep reading."

• Adding "why this data is being presented" before the graphs

• Structuring it to talk about the "process" and "challenges" rather than just the results

• Giving the words warmth through statements from employees and stakeholders

As a result, voices saying "I read it" emerged within the company, and the IR department also evaluated it as "a document we can use."

■ Summary of AI Tools and Prompts Used

Tool | Usage Scenario | Used Prompt (Excerpt) | NotebookLM | Structural analysis of reports and extraction of improvement points | "Is this document structured to reach the reader? Please organize improvement points structurally." | ChatGPT | Redesigning into a readable story structure | "Please make it a compelling structure following the flow of 'Challenge -> Initiative -> Realization -> Future'." | Perplexity | Comparison with other companies' reports and analysis of expression tone | "Please tell me examples of well-regarded, easy-to-read structures for ESG reports in the manufacturing industry."

■ Column: Who are you writing the report for?

ESG reports are documents seen not only by external investors but also by employees, business partners, and sometimes future job applicants.
When the goal becomes just "getting it out there," the "value of reading it" is lost.

Generative AI provides a perspective that transforms such documents into "communicated messages."
By simply adding a "narrative flow" and "reader intent" to the numbers and wording, a document becomes content.

Let's rethink with AI: "Is this document actually meant to be read?"

■ What to do in these situations?

Q: The ESG report is 200 pages long in PDF. It's not being read.
→ Use NotebookLM to organize the chapter structure and key points, and use ChatGPT to create and distribute a "digest story version."

Q: I don't know where it's effective to include employee voices.
→ Ask ChatGPT to suggest "places where adding a human voice would change the impression within this structure" for hints.

Q: Internal reviews aren't passing... How should I write to appeal to people who prioritize numbers?
→ Try using Claude to refine it into an "ESG story structure for the number-oriented demographic."

■ Reader Workshop: Thinking about an "ESG Report Restructuring Plan"

1. Organize information and diagnose the structure of existing reports with NotebookLM

2. Rewrite into a reader-centric story structure with ChatGPT

3. Compare with other companies' cases using Perplexity to find a presentation style unique to your company

4. Refine the cover and visuals with Canva to make it a "document people want to read"

■ Conclusion

An ESG report is not just a "report."
It is a "time to share" the company's promise for the future with the reader.

With the help of AI, transform the report from a "document to read" into "content to feel."
That small change gradually builds brand trust.

53 | Internal newsletters on Slack... Posting techniques common to teams that get reactions

(Chapter 6 | Cultivating brand "uniqueness" through words and culture)

■ Introduction: The loneliness of PR where "no one reacts"

Startup company "S-Form." About 80 employees.
Ai Terai from PR posted information once a week on the dedicated Slack channel "#internal-newsletter."Ai Terai was posting information once a week.

"Here is this week's PR coverage information!"
"A new member has joined!"

However, the post didn't even get a single stamp, and there were zero comments.
"Is anyone even looking at this...?"
Terai prepared the next post while feeling a sense of emptiness.

■ Case: Posts that generate reactions have a "design"

Terai used Perplexity to research "successful examples of using Slack for internal newsletters."
He extracted the common points of posts that get read.

• The amount of text is not too large

• There is a "hook" at the beginning that makes people want to read it

• The post includes "questions" or "choices"

Based on these insights, he input the following prompt into ChatGPT.

Prompt: "Please rewrite the following internal newsletter text for Slack to increase engagement. Please improve it by including a 'hook at the beginning,' using emojis and line breaks effectively, and designing questions."

Furthermore, he used Notion AI to organize "employee introduction templates" and prepare a stock of post ideas. Before posting to Slack, he also used
Claude to adjust the tone and writing style.

■ Solution: Things change just by adding "white space" and "questions" to posts

Terai realized that internal newsletters that get reactions have three things in common.

1. A design where people "read while reacting" rather than reacting after reading

2. Written as a "conversation" rather than just sharing facts

3. Includes questions that encourage readers to make it "personal"

Keeping these three points in mind, he created a structure template using ChatGPT. After a few posts, the number of reactions improved significantly, and Slack became flooded with comments like "I'm reading this!"

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt) | Perplexity | Researching examples from other companies | "Please tell me examples of post structures for internal newsletters that get reactions on Slack" | ChatGPT | Rewriting post text and template design | "Please rewrite this into a Slack post that increases engagement. Please be creative with questions, emojis, and white space" | Notion AI | Organizing employee introduction templates | "Please create a structure template for 'employee introduction articles' that can be used in internal newsletters" | Claude | Fine-tuning post tone | "Please make the tone casual yet respectful, suitable for Slack"

■ Column: Information that gets reactions has "white space"

Is your goal to "convey all the information"? Internal newsletters are read when they are designed as "conversation starters" rather than complete news pieces.

Generative AI is a partner that thinks with you about "what to say, in what order, and how to ask questions." A PR stance of "involving people" rather than just "delivering information" begins to work within Slack.

■ What should I do in these situations?

Q: I'm stuck in a rut, posting similar content every time
→ Create a 'list of post topic categories' in Notion AI and use ChatGPT to expand the angles of your content.

Q: My posts feel one-sided... I want to capture more employee voices
→ Create a 'template of questions to ask employees' in Claude and design interactions on Slack.

Q: Management doesn't understand the purpose of Slack posts
→ Have ChatGPT generate an 'explanation of the value of internal newsletter Slack posts' and use it in your reports.

■ Reader Workshop: Creating a 'Slack post template that generates engagement'

1. Collect examples of Slack posts from other companies using Perplexity

2. Generate a template structure suited to your company using ChatGPT

3. Organize post stocks and employee introduction topics in Notion

4. Adjust the tone of voice with Claude and post to Slack

■ Conclusion

Internal newsletters can evolve from 'information sharing' to 'organizational development'.
By designing for conversations to emerge within Slack, you can change the temperature of the company.

It might not be that there is no reaction, but rather a 'structure that makes it hard to react'.
Together with generative AI, design questions and space, and cultivate the 'culture of language' in your organization.

54 | Brand reconstruction phase, taking inventory of words that fit the 'current company'

(Chapter 6 | Cultivating brand 'uniqueness' through words and culture)

■ Introduction: 'I feel a sense of discomfort with our founding slogan'

Reflection, a B2B SaaS company in its 12th year.
The service is growing, and the team is expanding. However—

Shugo Yamauchi, the PR manager,felt a sense of discomfort with the brand's language.

'We are still using the catchphrase from when we were founded, but it feels a bit old.'
'When employees ask, "How do we want to be seen now?", I get stuck for an answer.'
'In new graduate recruitment, we've been told, "It was a bit different from what I imagined."'

He had a sense that the 'words were misaligned,' but he didn't know what or how to review them.

■ Case: Inventory and dialogue design to verbalize the 'current company'

First, Yamauchi uploaded past company slogans, CEO greetings, brochure copy, and past social media posts to NotebookLM and executed the following prompt.

Prompt: "Analyze the following set of words and list the 'keywords that are consistent as a brand' and 'expressions that may be out of sync with the times'."

Next, he designed the following workshop using ChatGPT.

Prompt: "Create a plan for an internal workshop to reconstruct 'brand language that fits the current company.' Please include the following perspectives: - Does it align with the current business phase? - Is it language that employees can speak about as their 'own'? - Is there any misalignment in the appeal to external audiences?"

Furthermore, he generated multiple brand keyword candidates using Genspark. He then used
Claude to translate them into easy-to-understand language for external audiences and refined the PR materials.

■ Solution: Brand 'words' can be reinvented

What Yamauchi realized was that 'brand language can be reconstructed not as a 'past definition' but as a 'current shared understanding'.'

• Updating language is not about denying the past, but about capturing the 'now'

• Extracting new keywords from employees' own words increases organizational unity

• AI adjusts the 'intersection' between the words put out to the world and the words that resonate internally

The new slogan is 'Making work more transparent.'
It has been well-received by employees and is beginning to be used in recruitment and sales settings.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Prompt Used (Excerpt) | NotebookLM | Extraction and analysis of existing brand language | 'Visualize keywords that appear consistently and expressions that are out of sync with the times' | ChatGPT | Internal workshop design | 'Design an internal WS to reconstruct brand language. Make it fit 'who we are now'' | Genspark | Brand keyword generation | 'Generate multiple brand slogan/tagline candidates that fit our current phase' | Claude | Translation for external use and tone adjustment | 'Adjust the language to a 'clear and trustworthy tone' for external audiences'

■ Column: A brand is a 'shared landscape created by words'

A brand does not start with 'design' or a 'logo,' but with 'words' that align on 'how we want to be.'
Words create the 'gravitational field' of an organization.

Generative AI is a presence that shows you both an inventory of the past and hypotheses for the future when updating those words.

'Who are we now?'
Brand reconstruction begins by facing this question again through words.

■ What to do in these situations?

Q: I feel hesitant to change the founder's words...
→ Use Claude to create a 'structure that connects to the founder's words' for a 'reconstruction that maintains respect.'

Q: What is our 'uniqueness'? It varies from employee to employee.
→ Aggregate internal documents and interviews in NotebookLM and extract the 'common language'.

Q: There is a discrepancy between external and internal messaging.
→ Have ChatGPT generate a 'comparison table of expressions for external vs. internal use' and tune it.

■ Reader Workshop: 4 Steps to Refine Your 'Current Brand Language'

1. Integrate and analyze past brand language, brochures, etc., using NotebookLM.

2. Design an internal workshop using ChatGPT.

3. Generate candidate slogans and taglines using Genspark.

4. Adjust tone and translate language for external communication using Claude.

■ Conclusion

Language is not something that 'must never be changed,' but rather something that 'can be changed when the time is right.'

Company phases, employees, and businesses all change.
Therefore, it is the job of PR to reclaim the 'language' that is appropriate for those changes.

Generative AI can be your partner in this, helping to articulate your uncertainties.
Let PR lead the way on the journey to put 'who we are now' into words.

55 | Alliance Announcements: The Secret to Designing Without Conflict Over 'Who is the Subject?'

(Chapter 6 | Cultivating Brand 'Identity' Through Language and Culture)

■ Introduction: 'We have a problem with our name being listed second'

Medical startup 'MediFuture' was set to announce a business alliance with a long-established pharmaceutical manufacturer.

PR representative Takahiro Sugiyama was struggling with his first alliance announcement.

• 'The meeting has stalled three times just over which company should be the subject of the press release...'

• 'I got an email from the other company's PR saying, "Please write our company name first"...'

• 'Every time we adjust the text, legal and sales departments also chime in with revisions.'

What was needed was the PR 'language design skill' to communicate as 'equal partners' while respecting each other's positions and benefits.

■ Case: Agreeing on Drafts Through 'Structure'

First, Sugiyama used ChatGPT to execute the following prompt and identify 'structural patterns for partnership announcements'.

Prompt: 'Please structure a draft press release for an equal business partnership. Create a format that avoids making either side appear superior while organizing the benefits, objectives, and background of both parties.'

The 'co-creation structure' release draft that was output looked like this:

• Start at the beginning with a sentence mentioning 'both companies jointly'

• A structure that alternates the 'background,' 'purpose,' and 'significance' as described by each company

• Include a paragraph emphasizing 'synergy through collaboration' from the middle onwards

Based on this structural plan, Claude was used to generate a 'tone natural for both sides.' Adjustments were made to parts where expressions were inconsistent or where one side's benefits appeared too prominent.

Furthermore, with visual materials created in Canva, consensus-building with management and legal departments went smoothly.

■ Solution: Creating consensus through 'dialogue design' rather than the 'order' of words

Sugiyama realized that the cause of friction was not 'whose name comes first,' but 'a lack of agreement on the structure.'

• Presenting the structure in a 'common format' makes wording adjustments constructive

• You can create a 'sense of fairness' through the story while respecting the other party's words

• You can redefine the 'subject' not by whose company it is, but by 'what we aim to achieve together'

Ultimately, the content was finished in a way that both companies described as a 'pleasant announcement.'

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Prompt Used (Excerpt) | ChatGPT | Extraction and organization of structural patterns | 'Please provide multiple 'fair structure' options for an equal business partnership release.' | Claude | Tone adjustment and balance check | 'Please refine the benefits and intentions of both parties into a natural style so that it does not lean toward one side.'

Canva | Creating materials for consensus-building | 'Please provide visual ideas to illustrate this partnership. Include benefits, role division, and future prospects.'

■ Column: For PR, the 'subject' is the design of a relationship

'Which is the subject?' may look like a battle over 'words,' but it is actually a battle over 'relationships.' What PR should design is not just the text, but the form of the relationship—'how we speak together.'

Generative AI is useful for adjusting the neutrality of the structure and the tone of the emotions. The level of refinement in words that make people and companies 'appear equal' becomes the intersection that creates the future.

■ What to do in these situations?

Q: The other company insists that their name must come first.
→ Use ChatGPT to generate alternative ways to arrange company names (e.g., alphabetical order, equal parallel representation) to use as negotiation material.

Q: The internal legal department pointed out that the wording is too vague.
→ Have Claude check for risky expressions from a 'legal perspective' and refine the language to be legally precise.

Q: The significance of the business alliance is difficult to convey in the first place.
→ Use NotebookLM to analyze past alliance cases and news articles to extract 'common success structures'.

■ Reader Workshop: Designing the Structure of an 'Alliance Announcement Release'

1. Generate multiple 'fair structural proposals' using ChatGPT

2. Refine the 'balance of writing style' using Claude

3. Create illustrative materials with Canva to explain to internal and external stakeholders

4. Learn 'success patterns' from past cases using NotebookLM

■ Conclusion

Conflicts over the 'subject' occur because the structure is ambiguous.
How you present the relationship starts not with the order of words, but with the design of how you speak together.

Generative AI connects different positions through 'structure' and puts the landing point of the dialogue into words.
As a PR professional, be the 'initial designer' who builds that bridge.

56 | Multilingual support: Why 'cultural adaptation' is required rather than just translation

(Chapter 6 | Cultivating brand 'uniqueness' through language and culture)

■ Introduction: 'This phrasing is too Japanese and doesn't make sense'

Future Nexus, an IT service company operating in five countries around the world.

Shinnosuke Yamano from the Public Relations DepartmentShinnosuke Yamano was struggling with global press releases.

• 'Directly translated English doesn't resonate with local media at all...'

• 'Expressions that work at headquarters have shifted nuances overseas'

• 'I end up writing things in the Japanese style that culturally don't need to be said'

After receiving feedback from local staff saying, "This release doesn't work locally," Yamano realized the necessity of "cultural adaptation" rather than just "translation."

■ Case: Understanding culture is necessary for translation that truly "conveys" meaning

First, Yamano used Perplexity to research "multilingual PR strategies for global companies."

Next, they uploaded past releases (Japanese original and English version) to NotebookLM and visualized cultural gaps using the following prompt.

Prompt: "Compare the following Japanese manuscript and English translation, and pick up parts where there are 'cultural nuance differences.' Identify expressions that might feel 'unnatural' to local readers and propose more natural alternatives."

After that, they used Claude to readjust the tone to match the business culture of each country. Example: Emphasizing "positive future-orientation" in the US, and "sincerity and sustainability" in Europe.

Finally, they used ChatGPT to generate a "list of expressions that readers might find jarring" and built a system to check for landmine words and elements requiring caution in advance.

■ Solution: To "convey" is to shift the "cultural coordinates"

What Yamano learned is that "language is a ship that carries culture."

• It is necessary to redesign messages using "local thought frames" rather than just translating them

• The perspective of prioritizing "local naturalness" over "accuracy" is important

• AI can be utilized not just for "translation," but as an "editor that bridges cultural differences"

As a result, the response rate for global releases improved, and trust with marketing and PR representatives in each country deepened.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Prompt Used (Excerpt) | Perplexity | Trend research for multilingual PR | "Please tell me about successful cases of multilingual PR strategies and cultural adaptation for global companies." | NotebookLM | Cultural gap analysis of past materials | "Please check if there are any cultural gaps in the translated expressions and propose natural paraphrasing." | Claude | Cultural adaptation of tone | "Please adjust this to a natural business tone for countries like the US, France, and India." | ChatGPT | Avoiding reader discomfort/pre-check | "Please generate a list of expressions that overseas readers might find jarring and propose alternatives to avoid them."

■ Column: Become a "cultural editor," not a translator

What is required in multilingual PR is not "literal translation" but "cultural editing."
Even if the language is the same, if the values are different, the meaning will not be conveyed.

Generative AI supports you in "delivering meaning" beyond words.
The role of PR is nothing less than building a "bridge" between cultures.

■ What to do in these situations?

Q: The text delivered by the translation company is too stiff and feels unnatural
→ Let's use Claude to refine it into a natural business tone, like a "native check."

Q: Local media pointed out that it 'doesn't make sense'.
→ Use NotebookLM to perform a comparative analysis of the original and translated texts to identify cultural gaps.

Q: The tone didn't match the local branch, and the release was rejected.
→ It is effective to create a 'local tone adjustment checklist' with ChatGPT and share it internally.

■ Reader Workshop: Tailoring a 'Cultural Adaptation Guide' for Your Company

1. Upload your company's translation track record to NotebookLM to analyze differences.

2. Create tone templates for each country using Claude.

3. Use ChatGPT to create a collection of examples of NG expressions and translation gaps.

4. Create and share a 'common checklist' with the internal local team.

■ Conclusion

Multilingual support is not just a translation task.
Public relations is being asked to play the role of a 'cultural translator' that redesigns 'who to communicate with and how'.

By using AI, you can learn about 'differences' and put 'consideration' into words.
PR professionals who can do that will become figures who are 'understood' all over the world.

57 | 'Involvement-style' PR techniques to turn employees' 'vague dissatisfaction' into charm

(Chapter 6 | Cultivating a brand's 'uniqueness' through words and culture)

■ Introduction: 'What is good about our company again?'

'Tokuma Seiko', a BtoB startup born from a long-established manufacturing company. It has about 90 employees.
It has grown rapidly in the last few years, and a PR department was newly established, but the challenge was 'internal enthusiasm'.

PR representative Tatsuya Fujii had heard these voices from within the company.

• 'It's not that I'm dissatisfied with the company, but there's nothing particularly worth bragging about...'

• 'I can't think of any points I'd dare to share.'

• 'There might be good points, but I can't put them into words.'

Everyone thinks it's a 'good company', but it hasn't been 'put into words'.
And, without being told anything, only the PR department kept moving.

■ Case: 'Vague dissatisfaction' is, when flipped, 'unverbalized charm'

Fujii first consulted ChatGPT as follows.

Prompt: "I want to derive positive value from employees' 'vague dissatisfaction.' Please read the free-text responses from the internal survey and visualize the underlying charm and points for improvement."

Furthermore, Fujii uploaded internal survey results, past employee interviews, and Slack chat logs to NotebookLM. Fujii gave the AI the following instructions.

Prompt: "From the employee voices below, extract 'elements that can be converted into charm' and propose three angles for transforming them into stories. Example: Simple -> Sincere, Steady -> Sense of security"

What emerged were 'reverse perspectives' such as 'Reserved = Sincere attitude' and 'Lack of challenge = A workplace where you can work with peace of mind.' Fujii organized these into a 'Participatory PR Prompt Collection' using
Genspark. Furthermore, Fujii distributed 'templates for employees to speak in their own words' via
Notion AI.

■ Solution: Designing for 'getting them to speak' ignites PR within the company

Fujii realized the importance of 'designing for getting them to speak.'

• Unorganized charm lies behind dissatisfaction and discomfort

• AI can discover that 'hidden value' and turn it into 'words'

• When more employees speak in their own words, PR shifts from 'outward messaging' to 'voices emerging from within'

Internal newsletters and employee-led communications gradually became more active, and eventually, applicants began saying, 'I could see the atmosphere of the workplace.'

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt)
ChatGPT | Positive conversion of dissatisfaction | 'Please reconstruct employees' "vague dissatisfaction" as positive charm'

NotebookLM | Analysis of internal statements and surveys | 'Extract "unverbalized charm" from employee statement logs'
Genspark | Templatizing employee-participatory PR | 'Output internal PR prompt templates that are easy for employees to talk about, categorized by job type and values'
Notion AI | Creating communication templates for employees | 'Generate writing templates that allow employees to talk about the company's charm in their own words'

■ Column: Dissatisfaction is a seed for the 'possibility of words'

"I feel somewhat uncomfortable," "But it's not worth mentioning."

Such voices might be 'signs' that the company is ready to move to the next phase.

By using generative AI, you can gain the 'power to put into words' voices that are often overlooked.
The role of public relations is to pick up the stories that exist within the 'silence.'

■ What to do in these situations?

Q: There are too few positive opinions in the internal survey
-> Upload the responses to NotebookLM and have ChatGPT convert them into 'potential value'.

Q: Employees won't cooperate with PR
→ Use Genspark to generate a "list of easy-to-discuss themes" or "prompt cards for easy storytelling" to encourage them to speak in their "own words."

Q: It's risky to share "complaints" externally as they are
→ It is effective to use Claude to adjust them into "positive transformations that maintain sincerity."

■ Reader Workshop: Creating a starting point for "Internal PR that gets people talking"

1. Visualize internal surveys and Slack comments using NotebookLM

2. Reconstruct them as "hidden charms" using ChatGPT

3. Prepare "prompts for easy storytelling" with Genspark

4. Share as templates for employees in Notion to foster a culture of spontaneous communication

■ Conclusion

It is many times more powerful for employees to speak than for the PR department to speak.
To draw out those voices, the perspective of "not denying complaints" and "translating them into charms" is essential.

Generative AI is the ultimate partner to help with that "translation work."
When employees' voices start to move, a "brand" begins to emerge from within the company.

58 | Prompt design to verbalize company identity through AI and employee workshops

(Chapter 6 | Cultivating brand "identity" through words and culture)

■ Introduction: "Who decides what our company is like?"

Ten years since its startup, with over 200 employees,Lifescope Inc.. As they transitioned from the venture feel of their founding days to a "mid-sized company," PR team member
Yui Tajima was worried.

• "In the past, 'speed' was our identity, but now it's not just that."

• "We have an MVV, but it's not being talked about as our 'identity' on the front lines."

• "There is inconsistency in the words coming from employees' mouths."

She tried many times to put "company identity" into words, but it always felt like it was missing the mark.

■ Case: Finding "true identity" from employees' "everyday language"

Tajima first uploaded past recruitment content, internal newsletters, Slack logs, and all-hands meeting minutes toNotebookLM. She used the following prompt to have the AI identify "common expressions and values."

Prompt: "From the following internal documents and employee statements, extract the 'commonly repeated words' and the 'underlying values,' and verbalize them as 'company identity.'"

Next, I utilized ChatGPT to convert the extracted words into a 'prompt for employee workshops.' The goal was to 'create a space for dialogue.'

Prompt: "Using these keywords as a starting point, generate 5 workshop prompts that will encourage employees to think and talk about 'our company's identity.'"

Furthermore, I prepared workshop materials using Canva and used Notion AI for real-time recording, summarization, and sharing of the discussions.

■ Solution: Verbalizing 'Identity' through 'Co-creation' rather than 'Excavation'

The answer Tajima arrived at was, 'A company's identity is not something decided by someone, but something discovered by everyone.'

• Extracting 'everyday language' with AI → Visualizing common values

• Designing 'mechanisms for dialogue' with ChatGPT → Replacing values with employees' own words

• Real-time organization with Notion → Establishing 'identity' as a shared understanding

After the company-wide workshop, employees naturally began talking about 'identity' on Slack. It was the moment when the words stopped being just empty concepts and began to take root as culture.
The words did not just stand alone; they began to take root as a culture.

■ Summary of AI Tools and Prompts Used

Tool | Usage Scenario | Used Prompt (Excerpt) | NotebookLM | Extracting common values from internal statements and documents | "Extract common expressions and values from statements and documents, and structure them as 'our company's identity'." | ChatGPT | Designing prompts for employee workshops | "Based on the extracted words, design questions and dialogue workshops that serve as a 'mechanism for dialogue'." | Canva | Preparing workshop materials | "Construct slides for a workshop where employees can think about our company's identity." | Notion AI | Summarizing and sharing workshop discussions | "Summarize this discussion content and organize it as a 'shared document'."

■ Column: 'Company Identity' resides in 'Employees' Catchphrases'

Even if you post a mission or MVV, it won't become an identity unless it falls into the 'words that come out of employees' mouths.'
True identity resides in the words of daily reports, Slack interactions, and casual conversations.

Generative AI is a partner that scoops up those 'unconscious words' and turns them into meaning.
The process of 'putting them into words together' is what allows culture to take root in an organization.

■ What to do in these situations?

Q: There are few people in the company who can 'verbalize,' so the workshop isn't progressing.
→ It is smoother to design 'questions that trigger memories' with ChatGPT and prepare statements that serve as hints in advance using NotebookLM.

Q: The philosophy and the language used on the front lines are far apart.
→ Use Claude to extract 'keywords that bridge' the philosophy and employee statements, and design a way to close the gap.

Q: Often, nothing sticks after a workshop...
→ Summarize the results in Notion and create a 'culture where you can look back and add to it at any time' to lead to continuity.

■ Reader Workshop: Design your own 'Company Identity Discovery Workshop'

1. Input internal statements and materials into NotebookLM to identify shared values

2. Generate "inquiry prompts" with ChatGPT that allow employees to speak for themselves

3. Create workshop slides in Canva and conduct them as a team

4. Summarize results in Notion to visualize and share them as "organizational language"

■ In Conclusion

A brand is not something you create, but something you nurture.
And the starting point for that is "what makes the company unique."

Creating a state where employees can express that "uniqueness" in their own words is the essence of public relations that fosters culture and trust.
Generative AI is a tool that translates those invisible values into "everyone's language."

59 | Techniques for creating mechanisms and "drawing out" stories to get employees to talk about their "favorite company"

(Chapter 6 | Nurturing brand "uniqueness" through language and culture)

■ Introduction: "The company's appeal? Hmm..."

BtoB SaaS company "LivComTech." 150 employees.
With numerous award wins and increasing media coverage, external attention was growing.

However, PR representative Gakuto Oshima had one concern.

• Even when asked for an interview, everyone says in unison, 'It's just normal.'

• There is little "employee posting" on social media, and the circle of empathy isn't spreading.

• The PR team intervenes too much in "getting them to talk," resulting in a stiff tone...

Even though he knew that "employees talking about the company" was the strongest form of PR, there was no mechanism to draw it out.

■ Case: Increasing "people who can talk" with "AI that draws things out"

Oshima first consulted ChatGPT.
He asked for examples of questions based on "empathy axes" that are easily discussed in employee interviews.

Prompt: "Please provide 10 interview prompts that can naturally draw out the moments and backgrounds where people feel 'I love this company.' Design them so that the respondent can realize 'good points they hadn't even noticed themselves.'"

Next, I used Claude to analyze past internal newsletters and interview articles.
I extracted common "emotional narratives" and "themes that are easy to talk about."

Furthermore, I created interview article formats for each department using Notion AI, and developed visual templates in Canva to make the "employees who speak" look appealing.

■ Solution: Creating a "space to speak" is where PR skills shine

It wasn't that employees didn't want to talk; they just lacked the "space and questions" to do so.
By designing this, the PR team's "ability to draw out stories" was recognized anew across the entire team.

• Prepare "questions that make people want to talk naturally" with ChatGPT

• Structure "tones and themes that are easy to talk about" with Claude

• Create templates in Notion so anyone can conduct interviews internally

• Visualize attractively with Canva → Activate sharing on internal social media

As the number of people who "speak about the company" increased internally, positive effects spread to recruitment, sales, and alliances.

■ Summary of AI tools and prompts used

Tool | Use Case | Prompt (Excerpt)

ChatGPT | Interview question design | "Please create 10 questions that allow employees to naturally talk about what they like about the company."
Claude | Discovering common "narrative axes" from past statements | "From the following statements, please structure 'ways of speaking and themes that evoke empathy'."
Notion AI | Interview article template creation | "Please create an internal interview article format for this department."
Canva | Visual design for showcasing employees | "Please create a visual template that fits this employee introduction article."

■ Column: Employees who can speak create a "favorite company"

A "company that employees talk about" resonates many times more than a "company that the PR representative talks about."

Getting people to talk is not about forcing them, but "drawing it out."
That design capability is the true value of PR.

Generative AI is a powerful assistant for designing the "questions" and "spaces" that naturally lead to employee narratives.
When the number of "people who can speak" increases, the company's "uniqueness" begins to spread outward.

■ What to do in these situations?

Q: Even when I interview them, the answers are always "cliché."
→ Use Claude to analyze past statements and start from "unexpected words" or "moments where emotion shows."

Q: I want to draw out words even from "employees who are not good at talking."
→ It is effective to create a "Yes/No interview format" with ChatGPT and have them open up starting from small questions.

Q: The article visuals are lackluster, and few people read them.
→ Use a "people-centric template" in Canva and focus on faces and comments; the impression will change significantly.

■ Reader Workshop: Designing a 'Narrative Mechanism' for Your Company

1. Create a 'Questionnaire to Encourage Storytelling' with ChatGPT

2. Organize 'Resonant Themes' from Past Narratives using Claude

3. Create Templates in Notion and Roll Out Interview Formats Company-wide

4. Prepare 'Employee-Centric' Visuals in Canva and Share Them Internally and Externally

■ Conclusion

A company where employees tell their own stories is the most powerful form of PR, generating empathy and trust.
To achieve this, designing 'ease of storytelling' is essential.

Generative AI is a powerful partner for designing questions, supporting the environment, and drawing out narratives.
The more a company's story is told, the more it spreads as something people want to 'champion'.

60 | Advancing the 'Re-verbalization' of Mission and Vision Through Dialogue with AI

(Chapter 6 | Cultivating Brand 'Authenticity' Through Words and Culture)

■ Introduction: The Gap Between 'Stated Words' and 'The Current Company'

NextArk, an IoT manufacturer based in a regional area, reached its 10th anniversary since founding.
The number of employees had increased year by year, and the business had expanded from smart homes to infrastructure.

Public Relations Manager Shota Sakuma was facing a certain challenge.

• The MVV created five years ago is no longer being discussed in the field

• Candidates in job interviews sometimes say, 'Honestly, I don't really get it'

• Management is cautious about changing it, and the process for redefinition is vague

'I want to update the words to fit the times while respecting the current ones.'
Faced with this dilemma, a 'dialogue-based review' using generative AI began.

■ Case: Taking Inventory of Words Through 'Dialogue with AI'

Sakuma first consolidated past MVV formulation materials, management documents, and employee interviews into NotebookLM.
To explore 'where the core of our values lies,' he input the following prompt.

Prompt: "Based on the following materials, extract our company's 'fundamental values.' At the same time, contrast the 'changes in the axis of our business' between the past and the present."

Next, I designed an internal survey using ChatGPT. I created a format where employees could freely write down 'keywords they resonate with now.'

Prompt: "In reviewing our mission and vision, I want to collect 'values that employees resonate with.' Please provide several question templates to elicit free-form responses."

Furthermore, I used Claude to categorize and aggregate the survey results, visualizing common keywords and differing perspectives.

Finally, I used Genspark to output multiple versions of 'drafts for re-verbalization,' which were then used in dialogues with the management team.

■ Solution: Designing a dialogue process of 'visualizing before deciding'

MVV is not something to be decided suddenly, but something that 'becomes visible.'

• Organize the 'past' with NotebookLM,

• Collect 'current voices' with ChatGPT,

• Analyze 'commonalities and differences' with Claude,

• Present 'multiple proposals' with Genspark.

By following this flow, we were able to design a process of 'carefully deciding while laying out multiple words' rather than 'deciding in a single word.'

As a result, we achieved a 're-verbalization' that both management and staff could strongly agree with, and the new MVV has begun to be used in internal training and recruitment materials.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Used Prompt (Excerpt) | NotebookLM | Inventory of MVV-related materials and linguistic assets | "Extract the core of our company's values and culture from the following materials." | ChatGPT | Employee survey design | "Please provide question ideas that allow employees to verbalize the 'value they feel in the company now' for the MVV review." | Claude | Categorization and organization of survey results | "Analyze common value keywords and their backgrounds from the following free-form responses." | Genspark | Outputting variations of re-verbalization drafts | "Please provide 10 patterns of rephrasing the MVV based on our company's values and direction."

■ Column: MVV is something that 'becomes visible,' not something to be 'decided'

Mission, vision, and value statements are not something handed down from above, but something that emerges from within.

Generative AI becomes a partner that supports that 'process of becoming visible.'

It can integrate diverse materials, convert feelings into language, enable comparisons, and allow for the simultaneous consideration of multiple proposals. By leveraging that power, you can create an organization that is 'convinced by its own words.'

■ What should I do in these situations?

Q: Management says 'MVV should not be changed,' so I'm stuck.
→ Use ChatGPT to design an internal survey on 'whether the current MVV is being communicated,' and visualize the results quantitatively and qualitatively before starting a dialogue.

Q: There are too many opinions, and I can't consolidate them.
→ Use Claude to categorize common and individual terms, and Genspark to generate multiple 'compromise proposals' to make organization easier.

Q: Employees aren't interested in the MVV.
→ Create an 'MVV page told in employees' own words' on Notion, and share it on Slack so it gets cited and spreads.

■ Reader Workshop: Try 'AI Dialogue Design for Re-linguaging'

1. Integrate the company's past MVV and internal documents using NotebookLM

2. Design a survey with ChatGPT and collect 'empathy words'

3. Organize the collected words with Claude and visualize overall trends

4. Generate multiple new MVV proposals with Genspark and discuss them as a team

■ Conclusion

Words are like an 'OS' that sometimes needs an update.
It is natural to adjust your Mission, Vision, and Values in line with growth and change.

If you proceed while engaging in dialogue with generative AI, you will surely find the optimal words to bridge your 'vision' and 'reality'.

61 | How to avoid failure in announcing a change of president — Methods for organizing expectations, reassurance, and the future

(Chapter 7 | What PR should do in times of emergency)

■ Introduction: Announcing a 'major event' is all about designing 'who to tell what'

BtoB SaaS company 'Sukimalink,' with about 120 employees. The resignation of the founding president and the appointment of a new representative had been officially decided, and PR representative
Kanta Yamazaki was nervous.

• Struggling with how to write the 'reason for resignation' in the press release

• Confused about the difference in tone for the media versus employees

• Lacking confidence in how to convey 'expectations for the future'

You must design not just the 'facts,' but the 'atmosphere.' Under that pressure, Yamazaki began to organize his thoughts using AI.

■ Case: Speaking to audiences with different perspectives using a common "structure"

First, Yamazaki used the following prompt in ChatGPT to extract a "template for the structure of a CEO succession announcement."

Prompt: "I want to create announcements for a CEO succession for shareholders, employees, and the media, respectively. Please provide a structural template based on 'Expectation, Reassurance, and Future'."

He realized that it was necessary to adjust the tone and information density for each target audience.

Next, he used Claude to import past examples from other companies and compare their sentence structures and wording.

Prompt: "Compare the following CEO succession press releases and analyze the differences between 'common phrasing' and 'effective communication.' Based on that, provide improvement suggestions for our company."

He also aggregated internal vision documents, IR materials, and past messages from the representative in NotebookLM to confirm the "consistency of the tone of voice."

■ Solution: Using a three-layer design to ensure "no one is left behind"

What Yamazaki arrived at was a three-layer message design.

• For shareholders: Management stability and future outlook (Logic + Track record + Numbers)

• For employees: Succession of corporate culture and values, dispelling anxiety (Reassurance + Empathy)

• For the media: Newsworthiness and storytelling (Change + Expectation)

While refining this framework with ChatGPT, he also created a visual relationship diagram in Canva. By designing it to tell the "story that begins with the succession" rather than just the "reason for the succession," he was able to realize an announcement that generated anticipation rather than backlash.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Prompt Used (Excerpt) | ChatGPT | Creating structural templates, message drafts | "Classify the CEO succession announcement structure into 'Expectation, Reassurance, and Future,' and adjust it for each reader." | Claude | Analyzing other companies' CEO succession announcements and proposing rephrasing | "Extract 'effective ways of speaking' from case analysis and create improvement suggestions tailored to our company." | NotebookLM | Analyzing the tone of internal documents, consistency check | "Check the tone consistency of this announcement by comparing it with past comments from the representative." | Canva | Creating relationship visuals | "Design an explanatory diagram for the CEO succession by stakeholder."

■ Column: Major announcements require both "reassurance" and "expectation"

A CEO succession is a "message to stakeholders" before it is "news." That is why it needs not only "no anxiety" but also "something to look forward to."

Generative AI connects the words of the past and the future and adjusts the tone for each reader. If you do not make a mistake in the "design of the words," the announcement of change becomes an opportunity.

■ What to do in cases like this?

Q: The CEO wants to keep the reason for the succession vague.
→ Have ChatGPT generate variations of "expressions that leave ambiguity while remaining convincing," and let them choose.

Q: The tone of voice differs between internal communications and IR materials.
→ Input both documents into NotebookLM to visualize the tone difference and gap, then have Claude propose a revised draft to smooth it out.

Q: There is anxiety among employees due to the change in president.
→ Collect employee feedback via internal chat and use Genspark to create an 'internal FAQ-style message' draft for a thoughtful response.

■ Reader Exercise: Practice in 'Three-Layer Design for Presidential Succession Messages'

1. Use ChatGPT to create a draft for a presidential succession message structured around 'Expectation, Reassurance, and Future'.

2. Use NotebookLM to compare with past materials and check for consistency in tone.

3. Use Claude to compare with other companies' releases and derive differentiated expressions.

4. Use Canva to create a 'Message Design Blueprint by Stakeholder' and share it internally.

■ Conclusion

In times of major change, it is important not only to 'convey correctly' but also to 'ensure it is received with peace of mind'.

By organizing your words with generative AI, you can create a design that reaches both the heart and the mind.
Even a presidential succession announcement can be turned into an opportunity through the design power of 'who to tell and how'.

62 | New Job Titles (like CPO): How do you explain a 'title no one has heard of'?

(Chapter 7 | What Should PR Do in Times of Need?)

■ Introduction: 'How do I explain this title to whom?'

Logistics SaaS venture 'Trazer' has newly established the position of CPO (Chief Product Officer).
Shota Izumifrom the PR department was puzzled by the announcement request from management.

• 'If the explanation is too wordy, it won't be read.'

• 'But just using the abbreviation won't convey who they are.'

• 'How do I turn the reason for creating this position into a story...'

A new job title carries more meaning than just a 'title'.
However, depending on how you express the background, its value can be halved or doubled.

■ Case: Talking about a job title through 'story' rather than 'function'

Izumi first researched the "definition and adoption status of the CPO role" using Perplexity. Next, she sent the following prompt to
Claude to summarize the job definition and social context.

Prompt: "Please summarize the general scope of a CPO (Chief Product Officer) role and why this position has been increasing in recent years, in a way that is easy to understand within the Japanese context. Then, reconstruct this into an 'easy-to-understand job title description' suitable for external PR."

Based on these results, she used ChatGPT to create a story linking the new CPO's persona with the company's growth context. She focused not on the reason 'why this title was added,' but on the necessity of 'why this person is taking on this role.'

Furthermore, she revamped the "company organizational map" using Canva. She created a layout where the positioning of the role could be understood visually.

■ Solution: It resonates when you speak of a "role = organizational intent"

Izumi structured the explanation not just as a simple job title description, but by including the vision of "why the organization considers this role important?" The points are as follows:

• A newly established role is a signal that the "center of gravity for organizational decision-making" is shifting

• Instead of just conveying the acronym, speaking in terms of "people + context" adds a human touch

• Showing the connection to the business phase increases persuasiveness

As a result, the media response to the new appointment release was positive, and understanding of the role also progressed within the company.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Prompt Used (Excerpt) | Perplexity | Background research on the CPO role | "Please tell me about the definition, necessity, and background of the CPO role." | Claude | Reconstructing social background and definition | "Please reconstruct the background of the increase in CPOs in recent years and their scope of duties for a Japanese audience." | ChatGPT | Building a story for the release | "Please describe the necessity of this person's appointment as CPO in line with the company's growth story." | Canva | Visual design of the organizational map | "Please create an organizational chart that clearly conveys the CPO position."

■ Column: A new role is a "symbol of change"

CPO, CHRO, CLO, CQO... Newly established titles are not just about organizing roles; they are messages about "what the organization has started to value."

Generative AI can not only organize definitions and roles but also polish language by layering "company intent" onto them.

Speak not as an explanation, but as an intention. That is the trick to turning a title no one has heard of into a "title people want to hear about."

■ What to do in these situations?

Q: The definition of CPO is vague, and explanations differ from person to person.
-> Use Claude to analyze CPO cases including those overseas, and use ChatGPT to create separate drafts for internal and external use.

Q: There are skeptical voices within the company asking, "Is this new role necessary?"
-> Use NotebookLM to reference management documents and past strategy documents to structure and show the basis and background for "why this role is needed now."

Q: If the explanation gets too long, it doesn't get read...
-> It is effective to create an "organizational evolution timeline" in Canva and supplement it with visuals that convey the story using only the key points.

■ Reader Workshop: "How to Communicate New Job Titles" Training

1. Create both "internal/external" job title explanations using ChatGPT

2. Organize job definitions and social context using Claude

3. Visualize the organizational position of the job title using Canva

4. Check for inconsistencies with past documents using NotebookLM

■ Conclusion

Communicating a new job title is also about putting the "future" of the organization into words.
By using generative AI, you can elevate your explanation beyond mere description into a "compelling narrative."

It is not just about listing names and titles,
but also conveying the "intent" behind those titles.
That is the job of public relations.

63 | Immediate judgment and preparation when stock prices plummet and media response cannot keep up

(Chapter 7 | What should PR do in an emergency?)

■ Introduction: "The stock price fell before the social media numbers settled down"

Sanjo Electric, a long-established domestic manufacturer.
Malpractice at a subsidiary was reported in an economic magazine, and the stock price plummeted by 8% that same morning.

PR Manager Takashi Matsunaga was caught in a whirlwind of responses in the hour before and after the disclosure.

• "The reports are ahead of us, and we cannot verify the facts."

• "Social media is starting to flare up before internal approval is granted."

• "IR, Legal, the President's Office... coordination is not keeping up."

What is fact, and where does speculation begin?
Even in such chaos, PR is asked to decide what to "release" and what "not to release."

■ Case: Immediate judgment by separating "three axes" with AI

Matsunaga immediately entered the following prompt into ChatGPT.

Prompt: "Based on the following report, please classify the information into facts, speculation, and evaluations. Also, for each element, please organize a list of 'what should be communicated now' and 'what should be sorted out later.'"

At the same time, NotebookLM was used to upload past press releases, governance-related documents, and IR materials.
It immediately checked for "consistency with previous communications."

Furthermore, using Claude, we generated three patterns of "expected public opinion, reporter questions, and shareholder reactions" based on the news report.
Public relations messages for each were immediately created using ChatGPT.

This clarified the criteria for deciding "what to release and in what order."

■ Solution: Release in the order of "Facts, Emotions, Future"

The decision-making principles Matsunaga derived are as follows:

1. Facts: "Concise information only on what has been confirmed at this moment."

2. Emotions: "Include at least a brief expression of gratitude for the concern caused."

3. Future: "Prioritize announcing the policy for investigation and recurrence prevention."

The emergency announcement material covering these three stages was formatted using Notion AI and passed through internal checks.

Three hours after the response, the situation calmed down. The next morning's stock price was up 1.2% from the previous day, partially recovering from the sharp drop.

■ Summary of AI Tools and Prompts Used

Tool | Usage Scenario | Used Prompt (Excerpt) | ChatGPT | Classification of news content and design of response priority | "Classify into facts, speculation, and evaluation, and identify what needs to be communicated now." | NotebookLM | Consistency check with past documents | "Check for contradictions with past disclosure materials and clearly state the differences." | Claude | Generation of expected questions and public opinion | "Generate patterns of reactions from reporters, shareholders, and SNS users to this report." | Notion AI | Editing of announcement materials | "Edit this PR message into an easy-to-read text and format it."

■ Column: In a crisis, "Subject, Tense, and Order" are vital

What is important when communicating information during confusion is the subject (whose words they are), tense (past, present, or future), and order (what to communicate first).

Generative AI calmly "structures" the information in front of you and shows you the order in which to communicate it.
Mechanical organization is effective precisely when emotions are wavering.

Behind calm writing, there is "organized thinking."
Behind organized thinking, there is "supporting AI."

■ What to do in these situations?

Q: Information is not yet confirmed, and I am in a state where I cannot say anything.
→ Use ChatGPT to output a "message limited to what can be communicated" and build a framework using only words that can be made public.

Q: Internal coordination is delayed, but the reporting is moving forward.
→ Use NotebookLM to check past disclosure materials and prepare at least the parts with differences so they can be shared first.

Q: I am afraid of questions being thrown at me during a press conference.
→ It is reassuring to have AI generate "expected reporter Q&A" with Claude and prepare answers with ChatGPT.

■ Reader Workshop: Designing the 'Axis for Immediate Judgment'

1. Try 'Deconstructing News Content' with ChatGPT

2. Extract 'Differences from Past Documents' with NotebookLM

3. Generate 'Question Patterns and Expected Answers' with Claude

4. Instantly Format 'Primary Announcement Materials' with Notion AI

■ Conclusion

What is tested in an emergency is not the 'ability to answer immediately,' but the 'quality of judgment.'

Rather than rushing to publish, take one minute to align your axis.
Generative AI is the strongest partner to support that 'alignment' phase.

Before putting words out, organize your thoughts.
Organized words create trust amidst chaos.

64 | Disaster Messaging: Designing Text that Balances 'Calmness and Warmth'

(Chapter 7 | What Should PR Do in an Emergency?)

■ Introduction: 'Feelings Run Ahead, Words Come to a Halt'

One morning, an earthquake with a seismic intensity of 6 occurred, centered in the Kanto region.
EduWave, an education startup with its headquarters in Tokyo, was also busy confirming the safety of its employees and responding to the situation.

The PR representative, Yuma Kawashima, was immediately tasked with drafting internal Slack messages and external announcements.

However, her hands stopped.

• 'If I don't show enough consideration for those affected, it could lead to a backlash.'

• 'But if the tone is too emotional, it will feel unnatural.'

• 'Balancing calm information with words that empathize with emotional distress is difficult.'

She had once learned in a PR training session that 'disaster messages need a temperature that is neither cold nor excessive.'
However, when she actually tried to write, she didn't know what the correct words were.

■ Case: Designing 'Emotion and Structure' Simultaneously with AI

Kawashima first entered the following prompt into ChatGPT.

Prompt: "Following the large-scale earthquake that occurred on [Date], I would like to create a message for both internal and external stakeholders. Please create a text that is both calm and warm, including the following elements: - Confirming the safety of employees and their families is the priority - Sympathy for those affected - Status of our company locations (no damage at this time/partially responding, etc.) - Future response policy and contact information"

The generated text had a perfect balance of polite tone and emotional 'range.' However, feeling that the writing was 'a bit stiff,' Kawashima sent the following additional prompt to
Claude.

Prompt: "Please adjust this text to be soft enough to convey a 'human touch.' Please use a tone that is not overly sentimental but helps to ease the reader's anxiety."

This revision added a gentle space to the expression.

Finally, Notion AI was used to format it for use in the internal newsletter, and the message was sent out on the same day.

■ Solution: A set of 'Information,' 'Empathy,' and 'Next Steps'

The lesson Kawashima learned is that 'peace of mind' is only achieved when the three elements of 'information to convey,' 'empathetic words,' and 'next steps (contact information)' are all present.

By using AI,

• comprehensive structure

• emotional tone adjustment

• ready-to-use formatting

became possible in a short time. "The more things shake, the stronger the 'core of words' becomes." It was a response that made that reality felt.

■ Summary of AI Tools Used and Prompts

Tool | Use Case | Prompt Used (Excerpt) | ChatGPT | Basic structure and draft creation during a disaster | "Please create a message for internal and external stakeholders following a disaster, with a calm and warm structure." | Claude | Tone adjustment and adding 'softness' | "Please edit this into a tone that is not too stiff, but not too light, and that is close to the reader's heart." | Notion AI | Deployment to internal/external formats | "Please reformat this text so it can be used in multiple formats such as Slack, internal newsletters, and the web."

■ Column: Especially during a disaster, let 'human voices' reside in your words

When a disaster occurs, what readers are looking for is 'calm information' and 'empathetic warmth.'

AI supports the former structurally and supplements the latter with editing tone.

Modern public relations involves not separating the 'design' of words from the 'emotion,' but supporting both. Can you produce a sentence that makes people think, 'This company has real people in it'? That is the first step toward trust.

■ What to do in situations like this?

Q: Should I prioritize internal or external communications?
→ Use NotebookLM to review past correspondence documents to organize priorities and workflows for both internal and external communications so you don't get lost.

Q: Should I release something even if the impact of the disaster is minimal?
→ Ask ChatGPT for "examples of messages when you're unsure whether to release." It is also effective to convey that there is no impact as a "source of reassurance."

Q: The tone varies within the company, and I can't make a decision.
→ Use Claude to generate drafts with multiple tones, and compare them to make an objective choice within the company.

■ Reader Workshop: Prepare a "Disaster Communication Template"

1. Create a framework for disaster response messages for your company using ChatGPT

2. Generate 2-3 patterns for tone adjustment using Claude

3. Organize it as a general-purpose template in Notion and share it internally

4. Regularly check for discrepancies with existing documents using NotebookLM and update accordingly

■ Conclusion

Communication during a disaster should be "fast, accurate, and kind."

AI can be a partner that supports all of these.
No matter how carefully you design it, it is meaningless if the communication is delayed.
Conversely, it would be counterproductive if it gives a cold impression, even if it is done in haste.

Calmness and kindness.
Let's borrow the power of AI to achieve both.

65 | Employee Health Initiatives: Techniques for Communicating the "Sentiment" Rather Than the "Policy"

(Chapter 7 | What Should PR Do in Times of Emergency?)

■ Introduction: "The policy is in place. But no one is talking about it."

Yosuke Morishita, who works in the healthcare promotion team at a major IT company, is a PR staff member in charge of the internal newsletter.Yosuke Morishitais a PR staff member in charge of the internal newsletter.

While collaborating with the industrial physician, HR, and labor teams, he introduced "stress care consultations" and "work-life balance support systems" as part of health management, but when it came time to communicate these policies to the company, the response was sluggish.

• "The policy is good. But it's not being conveyed as something that concerns them personally."

• "The hard work is ending up as just an 'announcement' on the company bulletin board."

• "If management's 'we're doing something' attitude is too prominent, it will be out of sync with the employees' perspective."

Morishita felt that what was needed was not an "explanation of the system," but an "articulation of the underlying sentiment."

■ Case: Executing the conversion from "System" to "Empathy" using AI

Morishita first sent the following prompt to ChatGPT.

Prompt: "Please restructure the content of the following health initiatives into a message that conveys not just an explanation of the system, but also 'what the company thinks of its employees.' Please create a structure that fosters empathy and understanding, and ensure the tone is not one-sided from management." [Examples] - Strengthening mental health consultation services - Work-life balance adjustment systems - Support for follow-up health checkups

ChatGPT refined the text into a "story-based format" that included not only an overview of the systems but also the background of their creation and the intent behind their introduction.

Next, Morishita requested a "restructured draft incorporating employee voices" using Claude.

Prompt: "Please insert short stories into this text, such as employee testimonials or the 'sense of security felt because the system exists,' and organize it into a structure that puts a human face on the initiative."

Furthermore, Morishita revamped the design of the system explanation slides using Canva.
They were able to transform the materials from "explaining the mechanics" to "materials that prioritize empathy."

■ Solution: From "Conveying" to "Speaking to"

Morishita realized that for health initiatives to be conveyed not just as a "system" but as "corporate sentiment," what was needed was not "design" but "conversation."

• System mechanics → From the perspective of management and HR

• Received feedback → From the perspective of employees

AI proved useful in the "design of the conversation" that connects these two sides.
As a result, the number of views on the internal newsletter tripled, and the utilization rate of health initiatives increased.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Prompt Used (Excerpt)
ChatGPT | Storytelling including the background of the system | "Please change the health initiatives into a story-based text that fosters empathy and understanding."
Claude | Structural editing adding employee voices | "Please insert short employee stories into this system introduction to create a structure that puts a human face on it."
Canva | Design formatting for internal newsletters and materials | "Please think of slide design ideas and catchphrases that fit this content."

■ Column: The "order of emotions" is key to introducing systems

When communicating health initiatives, if you start with the "system" itself, a sense of "obligation" inevitably seeps through.

However, simply by placing the background of why the system was created and the management's desire for employees to "continue working at this company" first, the way the reader opens their heart changes.

AI helps organize the order of emotions.
The era of "emotional design" is now necessary for system introductions as well.

■ What to do in situations like this?

Q: It's just a list of policy descriptions, and no one is reading them...
→ Try asking ChatGPT to 'reorganize this in an order that resonates with employees'.

Q: It's difficult to gather employee feedback.
→ You can use Claude to generate multiple hypothetical stories about 'how it feels to use the policy' and use them as source material.

Q: There is no consistency in the design of our internal newsletter.
→ It is effective to select a 'template for internal health initiatives' in Canva and standardize the format.

■ Reader Workshop: Designing 'Policy Introductions That Convey Your Intent' with AI

1. Reorganize initiatives into a 'story format' using ChatGPT

2. Supplement with feedback from an 'employee perspective' using Claude

3. Compile into visual materials using Canva and share internally

4. Check and unify consistency with other health initiatives using NotebookLM

■ Conclusion

Health initiatives should be communicated through 'intent' rather than 'policy'.

Can you make the people working there feel that 'the company is truly thinking about them'?
That is what leads to both policy utilization rates and corporate culture.

Generative AI becomes the 'translator of that intent'.
If you can refine words that move between structure and emotion, policies will change from 'notifications' to 'dialogues'.

66 | How to fuse 'science' and 'storytelling' to communicate research results to the public

(Chapter 7 | What should PR do in times of emergency?)

■ Prologue: There was a 'life' behind the mathematical formulas

MedicalPath, a startup collaborating with a medical university, succeeded in developing a groundbreaking testing technology.
They planned to issue a press release jointly with the university they had been conducting joint research with.

PR representative Takuto Ishikawa was concerned, even while paying attention to the accuracy of the content.

The research had a personal motivation from a researcher who wanted to save people suffering in clinical settings.
However, that was not reflected at all in the press release draft.

Ishikawa was torn between 'technical accuracy' and 'emotionally appealing storytelling'.

■ Case: Using AI to Transform Science into a Human Story

Ishikawa first sent the following prompt to Claude.

Prompt: "Reconstruct the following research results into a story format that the general public can understand. Please include the researcher's underlying thoughts and challenges in the clinical field, and write it in a way that allows readers to empathize with why this research is important." [Content] - AI model that detects early cancer risk from blood - Researcher: Professor Nakagawa of XX Medical University - Research motivation: Based on the experience of his own family member not being diagnosed early

Claude generated a draft that tells the results as a story centered on the researcher's motivation.

Next, he used ChatGPT to add technical explanations and balance the content.

Prompt: "Please supplement this story-style draft with accurate technical explanations using easy-to-understand analogies. Please also paraphrase technical terms and provide explanations that could be turned into diagrams."

Furthermore, he used Canva to create a 'one-page diagram' and a 'personal background introduction' to accompany the press release, making it easy to share with the media and within the company.

■ Solution: Stories Reveal the 'Contours of Science'

Ishikawa realized that 'what needs to be conveyed is not just the technology itself.'

• The 'motivation' behind the research

• The 'social issues' being addressed

• The researcher's 'personal perspective'

By weaving these into a story, he created a structure that reaches the reader's heart. As a result, the press release was picked up by multiple media outlets, and he received requests for lectures and articles.

■ Summary of AI Tools and Prompts Used

Tool | Usage Scenario | Prompt Used (Excerpt)
Claude | Reconstructing research results into a story format | "Please create a story article that includes the background and thoughts of the research, and is relatable to the general public."
ChatGPT | Supplementing technical explanations and paraphrasing terms | "Please explain this technology using analogies and a structure that can be understood by non-experts."
Canva | Creating diagrams and personal introductions to supplement the story | "Create a one-page diagram summarizing the flow of these research results and an introduction card that conveys the researcher's thoughts."

■ Column: An Era Where Science Also Needs 'Narrative'

Science used to be a world that competed on facts and data.

However, today, a 'context of empathy' is required to implement it in society. Stories serve as a tool to give complex technology a 'tangible contour.'

AI is also a capable storyteller for science. A new form of public relations is beginning, connecting scientific facts with the human heart.

■ What to do in situations like this?

Q: I don't know how to communicate research results to the general public.
→ Try asking Claude to 'tell the story of the researcher's motivation and results'.

Q: There are too many technical terms, and even people within the company say it's 'difficult'.
→ Ask ChatGPT for explanations using analogies or metaphors, and it will become much easier to understand.

Q: I don't know how to create diagrams.
→ Ask Canva to 'explain the mechanism of XX technology in one slide' and utilize visual templates.

■ Reader Workshop: Designing 'Story Materials to Communicate Research Results'

1. Turn research background and motivation into a story with Claude

2. Convert technical jargon into 'communicable language' with ChatGPT

3. Create a one-page diagram and researcher introduction card with Canva

4. Summarize similar communication examples with NotebookLM and confirm structural consistency

■ Conclusion

Research results are not just 'clumps of numbers'.
Within them lies a motivation, almost like a prayer, to save someone.

The job of public relations is to carefully weave that into a 'narrative'.
AI becomes a partner in that storytelling, acting as a bridge to deliver complex science to society in 'human language'.

67 | OSS Project PR: Find Common Ground with Engineers

(Chapter 7 | What Should PR Do in an Emergency?)

■ Introduction: Words Cannot Keep Up with the Codebase

The software development company 'CoreLayer' is currently developing multiple projects centered on OSS (Open Source Software).

Takahiro Nagai,Takahiro Nagaiwho is in charge of technical PR, has been worried lately.

There is public information and development stories about the OSS.
However, for the media and non-engineer users, it was an 'incomprehensible world'.

Nagai thought to himself.

■ Case: Placing AI in the “Gap” Between Dialogue and Translation

Nagai first uploaded the OSS GitHub repository, README, and developer notes to NotebookLM.

Using the following prompt, he extracted the main points and the value for readers.

Prompt: “Read the materials for the following OSS project and propose a ‘structure that clearly conveys its appeal’ to general engineers and corporate representatives considering its adoption. Reduce technical jargon and clarify the benefits.”

Based on the extracted keywords and use cases, he created an outline and tone proposal for a ‘technical introduction blog post’ using ChatGPT.

Furthermore, he requested the following from Claude:

Prompt: “Turn the ‘philosophy’ and ‘inspiration for development’ behind this OSS project into a story. Focus on ‘human motivations,’ such as the engineer’s passion and why this feature was created.”

→ By incorporating the ‘developer’s voice’ at the beginning of the article, the project’s relatability increased dramatically.

■ Solution: AI Excels at ‘Translating Philosophy and Language’

Nagai built a process with AI to convert ‘engineer-speak’ into a ‘common language.’

• Extract key points from code and materials using NotebookLM

• Format into drafts suitable for distribution forms like blogs and slides using ChatGPT

• Turn the ‘human drama behind the scenes’ into a story using Claude

As a result, page views for the OSS introduction page tripled.
External contributors to the project also increased, and the number of stars on GitHub continued to rise.

■ Summary of AI Tools and Prompts Used

Tool Usage Scenario Prompt Used (Excerpt)
NotebookLM Structuring and extracting key points from OSS materials “Analyze the OSS materials and create an appealing communication structure”
ChatGPT Formatting drafts for blog/slide structures “Reconstruct this technical information for a blog. Make it understandable for beginners”
Claude Storytelling the development motivation behind the technology “Turn the passion behind the OSS project into a story”

■ Column: Engineer’s Words Can Reach People If ‘Translated’

Many OSS projects present syntax and features first without answering the question, ‘Why was this technology born?’

However, there is usually ‘someone’s problem,’ an ‘ideal development environment,’ or a ‘challenge with peers’ behind it.

AI draws out and translates the ‘human motivations’ that are not spoken in code.
The key to OSS public relations is the transformation from ‘code to story.’

■ What should I do in these situations?

Q: The OSS documentation is too dry and hard to read.
→ Use NotebookLM to read the structure, and ChatGPT to add a narrative element.

Q: Engineers won't cooperate with PR.
→ Use Claude to create an interview structure based on their development motivations, and draw it out through simple interactions.

Q: The OSS introduction doesn't explain 'why it was created'.
→ Ask ChatGPT to 'write this as an interview-style article with the developer,' and it will naturally incorporate the 'story'.

■ Reader Workshop: Let's create an 'OSS PR translation template'

1. Have NotebookLM read OSS-related materials to extract common vocabulary and structures.

2. Reconstruct it as 'material for potential adopters' using ChatGPT.

3. Draft a story-based article using Claude that brings out the human drama behind the development.

4. Organize it as a template that can be deployed both internally and externally.

■ Conclusion

OSS PR is not just 'technical explanation that anyone can write.'
It is a field where you are tested on whether you can 'tell the story' of the developer's philosophy, motivations, and questions about the future.

AI acts as the bridge.
'Placing a story next to the code.'
That is the form of PR that gives life to OSS.

68 | How do you communicate a patent acquisition? The technique of converting 'technical appeal' into PR

(Chapter 7 | What should PR do in an emergency?)

■ Introduction: The PR wall that stops at 'It's amazing technology, but...'

IoT development company 'Reenect.' A B2B company with 150 employees and strong technical capabilities.

PR manager Ryota Mizuno was inwardly a bit troubled when he received the report of the patent acquisition.

The engineering team sent over technical documents and a summary filled with jargon.

However, it was unclear 'in what context,' 'with what words,' and 'to whom' it should be released to make an impact.

I was being rushed by the internal team asking, "Have you issued the press release yet?" but I was stuck and couldn't make any progress.

■ Case: Describing technology as a "vision of the future" rather than just "features"

Mizuno first uploaded the technical summary, development background memo, and competitor patent information to NotebookLM.

He organized the uniqueness of the technology and the PR perspective with the following prompt:

Prompt: "Read the following technical summary and extract the points where this technology is 'unique compared to other companies.' Then, summarize it according to the '3 elements to convey (social significance, potential for application, and track record)' for a PR document."

As a result, the following points became clear:

• It is an industry-first algorithm

• Specific use cases that reduce 'waste' in the field

• Future expansion into application areas (agriculture, logistics, medical care)

Mizuno then asked ChatGPT the following:

Prompt: "Reconstruct this patented technology into a PR document while depicting 'how future life and industry will change.' Please provide an angle and headline that reporters and users will find 'interesting.'"

Furthermore, he created visual materials with Canva and prepared a draft for a developer interview with Claude.

■ Solution: Expertise gains meaning by designing the 'order of communication'

What Mizuno discovered was not the 'technique for communicating difficult things,' but the importance of the 'order in which to communicate them clearly.'

• First, capture the reader's interest by talking about the 'vision of the future'

• Next, touch upon the 'context of solving problems'

• Finally, delve deeper with 'specific technical details'

The press release structured in this order was picked up by multiple professional and industry journals.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt) | NotebookLM | Structuring technical summaries and competitor comparison | "Summarize the unique points of this patent and the 3 key points for PR" | ChatGPT | Creating PR text from a future perspective | "Construct an easy-to-read PR document centered on how this technology will change society and life" | Claude | Creating a structure for developer interviews | "Create an interview article that draws out the motivation and struggles of development" | Canva | Creating illustrated materials | "Give me ideas for diagrams or flowcharts to convey this content"

■ Column: What Technical PR Needs Is Not 'Technology' but 'Translation'

Patents and new technologies are merely 'raw materials' for PR, not 'information' itself.

Depending on 'who you tell and how you tell it,' a simple report can transform into empathy, trust, or a hot topic.

Generative AI can be the ultimate training wheels for this translation work.

• Technical jargon → Empathy-driven language

• Technical details → Societal impact

• List of features → Future stories

Technical PR is a job that 'expands the world' through the power of editorial language.

■ What to do in these situations?

Q: The technical explanation of the patent is too difficult to understand.
→ Use NotebookLM to break it down, and then use ChatGPT to convert it into an 'explanation even a middle schooler can understand.'

Q: It is difficult to get cooperation from the development team.
→ Use Claude to create a 'template interview' and design questions that take little time, which will reduce their burden.

Q: The press release is too technical and isn't being read.
→ Ask ChatGPT to 'structure this story for non-engineers,' and the structure will change.

■ Reader Workshop: Designing the 'Order of Technical Communication' for Your Company

1. Load patent documents into NotebookLM and break them down into elements

2. Convert them into a 'non-expert-friendly structure' using ChatGPT

3. Turn developer voices into articles using Claude

4. Compile them into PR materials with easy-to-understand diagrams using Canva

■ Conclusion

Saying 'This is amazing technology' won't get through to people.

What you want to deliver is the context of 'why this technology is necessary for the future'.

Public relations is about organizing the order of communication.
If you can design that order with AI, technology becomes a story.

69 | Logo Refresh: What is the one sentence that prevents internal confusion and conveys the significance to the outside world?

(Chapter 7 | What should PR do in times of emergency?)

■ Introduction: I don't want it to end with just "It's changed"

Flowback is a manufacturing company celebrating its 25th anniversary. It has about 500 employees.

Along with the formulation of a new vision, it was decided to refresh the corporate logo.

PR Manager Makoto Taguchi was struggling with the text for the unveiling.

• "I want a message that conveys the background of the refresh."

• "But if it's too long, it won't be read."

• "I'm worried that people inside the company will say, 'The old one was better.'"

He received detailed concept materials from the designer, but if he used them as they were, it would end up being just a "discussion about appearance."

More than just "the logo is changing," he wants to convey "why it changed."

■ Case: Accepting change by "reconnecting" words

Taguchi first uploaded the logo change materials, past brand statements, and internal newsletters to NotebookLM.

He used the following prompt to organize the "common language" that should be conveyed both inside and outside the company.

Prompt: "Based on the following materials, decipher the background and story of the logo change, and extract 'three keywords necessary to gain understanding from inside and outside the company.' Then, create explanatory text in 100-character, 300-character, and 600-character patterns."

The keywords obtained were:

• Change (expansion of business domain)

• Inheritance (respect for founding spirit)

• Future-oriented (stance on environmental and social issues)

Based on this, I used ChatGPT to simultaneously create a 'draft for the internal newsletter' and a 'draft for the external press release'.

Furthermore, I used Canva to create before-and-after images of the logo refresh and diagrams aligned with the vision.

■ Solution: Messages are meant to build 'connections' rather than just being 'correct'

Taguchi realized that the goal was not just to change the logo, but that it was important to
retell 'who we are' through the logo.

• For internal staff, I used words that emphasized our 'connection to the past'.

• For external audiences, I used words that concisely conveyed our 'future stance'.

By adjusting the 'resolution of the message' for each purpose, backlash and confusion were kept to a minimum.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt) | NotebookLM | Organizing logo refresh materials and extracting keywords | 'Extract key keywords to communicate internally and externally, and output explanations for each purpose in 100, 300, and 600 characters.' | ChatGPT | Creating internal newsletters and press releases | 'Based on the background of the logo refresh, create messages for internal and external audiences for each purpose.' | Canva | Creating visual materials | 'Compare the Before/After of the logo change and suggest design ideas that convey the meaning.'

■ Column: Start brand changes with consideration for those who are 'drifting away'

When refreshing a logo or changing a vision, explaining it to 'those who cannot keep up with the change' is more important than anything else.

• Why change it now?

• Where has it changed?

• What remains unchanged?

Carefully delivering these three points builds trust in the change.

Generative AI is a strong ally that can tailor the optimal tone for 'different purposes and different people'. Instead of 'correct words', use 'words that reconnect'.

■ What to do in these situations?

Q: I heard voices from within the company saying, 'I liked the previous logo better'.
→ Use NotebookLM to check past philosophies and statements, and articulate the 'unchanging passion'.

Q: I cannot write short but essential sentences.
→ You can organize your thoughts by asking ChatGPT to 'narrow down the elements and write variations of 100, 300, and 600 characters'.

Q: I want to convey the reason for the logo change visually.
→ Using Canva to illustrate the meaning of the change (differences in shape, color, and symbolism) makes it easier to understand.

■ Reader Workshop: Articulating the 'Significance of Logo Changes' by Purpose

1. Upload logo change materials and brand documents to NotebookLM

2. Ask ChatGPT to draft text for different purposes (internal newsletters, IR, website)

3. Summarize the 'visualization of meaning' as a diagram in Canva

4. Review the created materials with stakeholders to unify the tone for internal communication

■ Conclusion

Refreshing a logo is an act of connecting the past and the future with the words of 'today'.

By using generative AI, you can simulate countless ways to craft 'accurate, concise, and impactful words'.

Why are we changing now?—that single sentence reflects the 'current position' of the organization.

70 | Legal and Governance Reform: Leveraging AI for 'Communicating Without Misunderstanding'

(Chapter 7 | What Should PR Do in Times of Crisis?)

■ Introduction: Why 'Talks About Rules' Fail to Get Through

Toa Materials is a mid-sized company expanding globally, primarily in the manufacturing sector, with 1,200 employees.

Under a policy to strengthen internal controls and corporate governance, new rules, systems, and guidelines were introduced all at once.

PR Manager Shota Fujii was tasked with summarizing the content for internal newsletters and intranet articles.

However, when he tried to write the draft—

• 'It's full of jargon and hard to read'

• 'It doesn't clearly convey which departments are affected'

• 'I was told the 'significance of the reform' wasn't written'

While accuracy was guaranteed with the help of the internal legal team and external lawyers, the barrier of 'not getting through' remained.

■ Case: Translating into a 'reader-centric' perspective while maintaining accuracy

Fujii first input internal documents (purpose of governance reform, summary of revisions, and background) into Claude and attempted to restructure them using the following prompt.

Prompt: "Please rephrase the following text in simple terms so that even non-specialist employees can understand it. Please restructure it, focusing particularly on 'why this rule is necessary' and 'what impact it will have on the front lines.'"

Furthermore, he created a draft for an intranet article using ChatGPT with the following prompt.

Prompt: "Please write an intranet article for internal use that conveys the overall picture and significance of the legal and governance reform. Please include concrete examples related to work, background, and future steps, in about 1500 characters."

Once the content was clarified, he created slide diagrams using Canva. By visualizing the background of the system introduction, the impact on the front lines, and the benefits, the material changed from something to be read to something that can be 'understood by seeing.'

■ Solution: 'Difficult' creates 'misunderstanding' — Balancing correctness and clarity

The biggest pitfall in public relations documents is being 'correct but not understood.'

Through rephrasing and visualization using AI, Fujii arrived at the following points:

• The order of explanation should be 'Why → How it changes → Does it relate to me?'

• Simplify technical terms as much as possible, or supplement them with footnotes and diagrams

• The tone should be conscious of 'dialogue' rather than 'rules'

As a result, comments on the intranet article doubled.
The correct answer rate for the simple quiz measuring internal understanding also improved significantly compared to the previous time.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Prompt Used (Excerpt)
Claude | Easy rephrasing of legal documents | "Rewrite in simple language for non-specialists, and restructure focusing on 'why it is necessary'"
ChatGPT | Overall structure and writing of intranet articles | "Structure an internal article conveying system reform in 1500 characters, including concrete examples and future steps"
Canva | Diagram and slide creation for internal materials | "I want to visualize the intent and impact of system changes for internal use on one page. Please propose a diagram"

■ Column: Text should be a 'guide to understanding,' not an 'explanation'

Topics like system reform, law, and governance often 'make people lose the will to read before they even start.'

That is why it is important to create a 'path that makes people want to read.'

The tone of the introduction, the order of the structure, the simplicity of the writing style, and the reality of the examples.

Generative AI can be used as a 'translator' to optimize these without changing the content.

Designing content tailored to the reader—that is the key to preventing misunderstandings and generating empathy.

■ What to do in these situations?

Q: The subject matter is too technical, and I can't even understand it myself.
→ You can organize your thoughts by using Claude to 'summarize, provide background, and convert into simple language' before planning the structure.

Q: The text is boring because it's just a list of policy names.
→ Simply instructing ChatGPT to 'add three concrete examples' will make it much more relatable.

Q: The response to intranet articles is weak.
→ Adding 'one-page easy-to-understand diagrams' created in Canva will make them more likely to be read.

■ Reader Workshop: 'Workshop on turning policy explanations into effective communication'

1. Have Claude read internal documents and restructure them to be easier to understand.

2. Generate multiple drafts of intranet articles using ChatGPT (1500 characters, 800 characters, summaries, etc.).

3. Design visual materials in Canva and post them together on the company intranet.

4. Design a 'one-click test' to check reader understanding (true/false tests or surveys).

■ Conclusion

Changing a policy and communicating it are two different things.

When the 'correctness of a policy' is not communicated, organizational trust is also shaken.

Generative AI is the best partner for delivering information 'clearly, accurately, and without misunderstanding'.

In the future, public relations will require the ability to design policy explanations as a 'story' as well.

71 | Responding to legal revisions: How to write for PR when 'not being read' is not an option

(Chapter 8 | Fulfilling accountability to systems, management, and society)

■ Introduction: An era where 'I just sent a notification' is no longer acceptable

Mirai Department Store, a mid-sized retail company, operates 80 stores nationwide.

What PR representative Yusuke Ohno faced was the internal announcement and external explanation of the 'Revised Labor Standards Act' in the spring of 2025.

The notification provided by the legal department is detailed and accurate, but it is difficult to read and hard to understand.

• "In the end, what is changing and how?"

• "How does this relate to the store floor?"

• "How should we explain this to customers and business partners?"

I don't want to just 'circulate a PDF and call it a day.' However, it is far too complex to explain clearly.

■ Case: Turning legal amendment information into text that people read as 'their own business'

Ono first used ChatGPT to execute the following prompt and create a draft for an internal intranet article.

Prompt: "Based on the following legal amendment details, please create an intranet article for store staff. Please summarize it in about 1500 characters, focusing on 'what is changing' and 'points to be careful about on the floor.' Please use a gentle and friendly tone so that even non-specialists can read it."

Next, he used Claude to adjust the tone of the 'announcement for customers and business partners.'

Prompt: "Please create an external explanation for this legal amendment. Please use a polite tone and focus on a sense of 'reassurance' so that customers and business partners do not feel anxious."

Finally, he used Canva to create a 'one-page slide for understanding legal amendment compliance' and posted it at the top of the company intranet.
He created a system not just to 'inform' about institutional changes, but to 'ensure understanding.'

■ Solution: If you decide 'who, why, and to what extent,' the text becomes clear

What Ono ultimately arrived at was that 'clarifying who you are communicating to also organizes what you need to write.'

• Store staff: What is changing + points for on-site response

• Customers/Business partners: Relieving anxiety + stance on response

• Entire company: Background and company policy

By using different AI tools for each target, he was able to adjust the 'temperature' and 'depth' of the text.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt)
ChatGPT | Creating intranet articles for store staff | "Please write an article explaining the Labor Standards Act amendment in a gentle tone, covering 'what is changing' and 'on-site points'"
Claude | Adjusting external explanations | "Please write an explanation of the legal amendment for external use that conveys a sense of reassurance"
Canva | Creating slides for legal amendment details | "Please visualize the key points of this legal amendment in a one-page slide for internal explanation"

■ Column: Legal documents have 'zero temperature,' so it is the job of PR to provide that warmth

Legal documents are calm and objective, which often leaves the 'reader's emotions' behind.

What PR needs to do is add 'warmth' to them.

For example—
・Why was this policy created?
・What step is this toward the future?
・What sense of security or value does this change bring?

AI can write accurately. However, to provide 'warmth,' the PR professional's own sensibility is required.

■ What should you do in these situations?

Q: It is difficult to switch tones between internal and external communications.
→ You can easily switch by asking Claude to 'please convert this into a tone for [target audience].'

Q: The content of the legal revision is too difficult to understand in the first place.
→ If you ask ChatGPT to 'please summarize this so that even a high school student can understand it,' your perspective will be clarified.

Q: My intranet articles often go unread.
→ If you turn them into slides using Canva and post them visually on the intranet homepage, the 'visibility rate' will increase.

■ Reader Workshop: Creating a 'Legal Revision Explanation Pattern' That Suits Your Company

1. Have ChatGPT create a draft of a 'key point summary article for company staff.'

2. Have Claude convert it into 'text for customers and business partners' and adjust the tone.

3. Create a 'one-page document' using Canva and use it for the intranet or as an email attachment.

4. By turning each format into a template, you make it easier to handle future requests.

■ Conclusion

Announcing policies carries the 'risk of not being read.'

If they are not read, it creates misunderstandings, delays on-site responses, and damages trust both inside and outside the company.

That is why you must be particular about 'how' you deliver your text.
Your strongest partner for this is generative AI.

Communicating a policy is also about showing the company's stance.
We are entering an era where that stance resides in your words.

72 | Inundated with interview requests and inquiries! Criteria for decision-making when you cannot keep up with media relations

(Chapter 8 | Fulfilling Accountability to Systems, Management, and Society)

■ Introduction: Sudden Attention, Too Much for a Solo PR Professional

The new AI educational material released by the education startup "EduBiz" went viral in an unexpected way.

It spread across social media, and inquiries from newspapers, television, and web media flooded in all at once.
However, the person in charge of PR is only solo PR professional Sho Kamiya.

• "The phone and emails won't stop..."

• "I can't keep asking management for approval on every interview request."

• "While I'm busy responding, I'm losing the ability to communicate what we actually need to say."

Even though "getting attention" should be a good thing, Kamiya felt overwhelmed and confused.

■ Case: Using AI to "Parallelize" Situational Judgment and Clarify Priorities

Kamiya first used ChatGPT to input the following prompt and create a "Priority Judgment Matrix" for handling various media.

Prompt: "Using the following media list as a reference, classify the priority into 3 levels based on 'social influence x compatibility with corporate strategy.' Then, propose a response policy (respond / put on hold / decline)."

By comparing media influence, past coverage examples, and corporate policy, he organized what to respond to and what to decline.

Next, he prepared template reply emails for each media outlet using Claude.

Prompt: "Create a polite email template to decline inquiries from media outlets like the following. Start with 'Thank you for your interest' and gently convey the reason for 'declining this time.'"

Furthermore, he uploaded past PR response records to NotebookLM to extract successful patterns from similar cases and use them as a reference for priority handling.

■ Solution: Creating a "Judgment Axis" with AI Leads to Calmness

By "outsourcing" the priority judgment to AI, Kamiya was able to organize his own thoughts.

• "What to do first"

• "Where to send a decline"

• "What primary information about the company should be communicated now"

Having these decision criteria transformed him from a "PR person who is just busy" into a "PR person who can make strategic choices."

■ Summary of AI Tools and Prompts Used

Tool | Use Case | Prompt (Excerpt)
ChatGPT | Creating criteria for prioritization | "Classify the media list by 'influence x strategic fit' and propose a response policy."
Claude | Drafting polite decline emails to media | "Create a PR-style email template for declining requests that conveys the reason politely."
NotebookLM | Learning from past interview responses and referencing decision criteria | "Extract the 'common factors of successful interview responses' from the following past response history."

■ Column: A 'Haste Makes Waste' Approach to PR Decision-Making

Panicking and reacting to everything because you feel you 'must respond' is counterproductive in the long run.

Instead of 'responding to every voice,' focus on the 'people you want to reach right now.'

Generative AI can calmly support that 'line-drawing.'
It allows you to make decisions with an exquisite sense of balance—neither swayed by emotion nor overly cold.

'If you're in a hurry, take the long way around with AI'—
That mindset might be exactly what PR needs today.

■ What should I do in these situations?

Q: I don't know how to determine media priority.
→ Use ChatGPT to create 'classification criteria for interview targets that fit your corporate strategy' to speed up decision-making.

Q: I spend too much time worrying about the wording when declining requests.
→ Prepare several 'decline email templates' with Claude so you can respond immediately.

Q: I get blamed by internal staff when I can't keep up with responses...
→ Summarize past response records in NotebookLM and output 'situation visualization materials' to make it easier to explain.

■ Reader Workshop: Let's Create an 'Interview Response & Decision Matrix'

1. Organize your 'priority media list for your company' using ChatGPT.

2. Prepare multiple 'polite decline email templates' using Claude.

3. Organize past interview response records in NotebookLM and extract successful patterns.

4. Share this matrix and set of templates within the PR team as an 'Emergency Response Pack.'

■ Conclusion

Media relations is an era where 'selection' is more important than 'volume.'

Within limited time and human resources, 'what to strategically pick up and what to let go.'
The essence of PR is revealed in that very decision.

Generative AI is a partner that brings us calmness and a higher perspective.
From 'PR overwhelmed by work' to 'PR that makes choices.'
Why not take that first step together with AI?

73 | What is the PR perspective that transforms board meeting materials into "materials that get read"?

(Chapter 8 | Fulfilling Accountability to Systems, Management, and Society)

■ Introduction: The uneasy feeling that "the materials aren't being read"

BtoB SaaS company "Security Cloud." 180 employees. With an eye on going public, initiatives were underway to improve the quality of management's decision-making.

PR Manager Ryota Tsujimura felt a sense of unease every time he looked through the monthly board meeting materials.

• "The numbers are there. But there's no explanation as to why these numbers are important."

• "The slides are there, but I don't know where I should be focusing."

• "I feel like the management team is 'making decisions based on vibes.'"

Realizing that the design of the materials was leaning more toward "stuffing them full" than "conveying information," Tsujimura began exploring approaches he could take as a PR professional.

■ Case: Giving the materials a "flow" and a "narrative"

Tsujimura first uploaded the past board meeting materials all at once to NotebookLM.

He analyzed the consistency, duplication, and omissions in the structure using the following prompt.

Prompt: "Read the following board meeting materials and evaluate the consistency of the structure, the narrative quality, and any duplications or omissions. Use 'how well the information necessary for management decisions is organized' as your judgment criteria."

Next, he created a material storyboard that aligned with the "management perspective" using ChatGPT.

Prompt: "Rewrite the following business progress report into a structure that management can immediately understand and make decisions on. Organize it in the order of Premise -> Current Status -> Points Requiring Judgment -> Proposal, and clearly state the title, summary, and points of focus."

Furthermore, he prepared visually easy-to-understand templates in Canva, and added supplementary comments to important slides using Claude.

■ Solution: Deepening thought by reducing the "burden of reading"

Tsujimura realized that understanding of the materials improved not by "reducing the amount of information," but by organizing the "order in which the meaning is received and how it is presented."

• PR perspective = Thoroughly considering 'how the recipient sees it'

• PR weapons = The order of communication, the tone of the words, and visual design

Documents that 'get read' are different from documents that 'make people read'.
Optimizing both structure and language to support decision-making—
that was the answer Tsujimura arrived at.

■ Summary of AI Tools Used and Prompts

Tool | Use Case | Prompt (Excerpt)
NotebookLM | Analyzing the structure of past documents | "Evaluate whether this set of documents has a consistent story and contains the elements necessary for management decisions."
ChatGPT | Storyboarding and structural rewriting | "Structure this in the order of Premise -> Current Situation -> Judgment -> Proposal, and make it a document that is easy for management to make decisions on."
Canva | Creating visual templates | "Create a template proposal that makes this document easy for management to read."
Claude | Generating supplementary comments for slides | "Add a one-sentence 'key point to convey' to the important slides."

■ Column: Documents also need 'narration'

A document is a 'presentation that doesn't speak'.

That is precisely why the order of words, emphasis, supplementary information, and questions—
everything must be designed.

Generative AI is good at refining the 'tone' of text.
It can do more than just organize information; it can evolve documents into ones that 'convey as if speaking'.

Documents without narration are not understood.
Documents also require the ingenuity of a 'narrator'.

■ What should I do in these situations?

Q: Board meeting documents are long and no one reads them properly.
→ Use NotebookLM to identify 'parts to cut' and 'duplicate parts', and use ChatGPT to redesign the structure.

Q: The story of the document is disjointed and doesn't connect.
→ You can organize it by using ChatGPT to convert it into a 'narrative-style' or 'decision-making-style' storyboard.

Q: The visual design is plain and doesn't get the point across.
→ Use a 'management meeting template' in Canva and work on the design of emphasized areas and white space.

■ Reader Workshop: Try 'Story-style Documents' at your company

1. Analyze past meeting documents with NotebookLM to find structural trends.

2. Rewrite into a 'decision-support structure' using ChatGPT.

3. Convert to a visually easy-to-read format with Canva.

4. Add supplementary comments and annotations with Claude to evolve it into a 'speaking document'.

■ Finally

Whether it gets read or not depends entirely on the quality of the document.

Leverage the PR perspective to embed 'design for communication' into your documents.
That will increase the speed of management's understanding and decision-making, and accelerate the organization's movement.

The quality of your materials is the very measure of your company's decision-making power—
Believing this, let's create "materials that speak" together with AI.

74 | Year-end report: Redesigning the structure to convey "gratitude" and "will"

(Chapter 8 | Fulfilling accountability to systems, management, and society)

■ Introduction: "Is your report just ending as a 'task'?"

Edrink is an education-focused startup headquartered in the Kanto region.
With 50 employees, they were at a milestone, entering their fifth year since founding.

Public relations representative Tatsuya Ohashi was tasked with creating the year-end report (a report for the entire company).
However, looking at past reports, he felt a sense of unease.

• "Just numbers and performance. There's no emotion."

• "Distributed as a PDF and that's it. No one has any feedback."

• "Next term's policy looks like nothing more than 'symbols'."

Couldn't this be a "document of gratitude and determination" rather than a list of numbers?
Thinking this, Ohashi decided to use AI to fundamentally rethink the structure of the materials.

■ Case: Designing a "narrative" for both "looking back" and "the future"

First, he uploaded the past five years of reports to NotebookLM to analyze the structure from the reader's perspective.

Prompt: "Read the following year-end reports and evaluate whether 'gratitude' and 'will for the future' are clearly conveyed. If they are not, suggest where and what kind of additions or restructuring are needed."

Next, he used the following prompt with ChatGPT to restructure it into a "story-style year-end report."

Prompt: "Please restructure this year-end report into a format that 'conveys gratitude to stakeholders and shows the will for the next term.' Suggest headings and text so that readers can 'be convinced of this year's hard work and feel excited about the next term.'"

Furthermore, he utilized "story-style slide" templates in Canva to change the structure to be visually softer and more memorable.

Claude was used to adjust the tone of the CEO's comments to a perfect balance of "not too stiff, but not too light."

■ Solution: When you turn materials into a "dialogue," the atmosphere begins to shift

The completed report became not just a report, but a "trigger for dialogue."

• Received feedback from employees and partner companies that it was a great opportunity to look back on the year

• The CEO's comments conveyed a sense of 'humanity,' which helped bridge the distance with new shareholders

• The goals for the next term were imbued with 'warmth of language,' and are now being quoted in internal notices and morning meetings

A 'retrospective' is not just a task.
Ohashi realized that it is an act of expression that delivers 'warmth' to the entire organization.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Prompt (Excerpt)
NotebookLM | Analysis of past reports | 'Does this express "gratitude" and "will"? If not, please suggest how to improve the structure.'
ChatGPT | Redesigning reports with a narrative structure | 'Please create a structure that conveys a retrospective + gratitude + will for the future. Include headings and a tone of voice.'
Canva | Creating warm visual templates | 'Please suggest a soft slide template suitable for a year-end gratitude report.'
Claude | Refining CEO comments | 'Please rewrite this comment in a tone that is "engaging but not too light."'

■ Column: Annual reports should leave behind 'memories,' not just 'records'

Annual documents are not just records.
The key is whether there are words that remain with the reader.

Numbers are important, but even more so are 'what we felt this year' and 'what we overcame together.'
Sharing such emotions builds the 'trust' that creates the future together.

Generative AI has the power to translate numbers and achievements into human language.
Let's borrow that power to deliver memorable messages to the organization.

■ What to do in these situations?

Q: The CEO's comments end up being the same every time
→ Ask Claude: 'Based on this term's highlights, please provide a draft comment with a personal tone.'

Q: The business report is too stiff and no one reads it
→ Try using this prompt in ChatGPT: 'Create a structure that employees and business partners will find convincing and relatable.'

Q: The slide deck is boring and doesn't get people excited
→ Utilize Canva's 'Thank You Report' or 'Narrative' templates to change it into a story-driven structure.

■ Reader Workshop: Applying the 'Gratitude + Will' structure to your company

1. Analyze past annual documents with NotebookLM to grasp trends in tone and structure

2. Try generating a 'story-driven' structure proposal with ChatGPT

3. Design warm slides in Canva and distribute them internally and externally

4. Polish CEO comments and messages with Claude

■ Conclusion

The year-end report is a device to ensure the end of the year is not just marked by a 'symbol'.

Words of gratitude become a force that illuminates the future.
Let's convey next term's policy through 'thoughts' rather than numbers.

Generative AI becomes a partner in finding and polishing those words.
To make your organization's 'story of the year' something that remains in someone's heart.

75 | M&A Announcement: How to avoid misjudging 'who the message is for'

(Chapter 8 | Fulfilling accountability to systems, management, and society)

■ Prologue: 'Somehow, everyone has gone silent'

The medical startup 'Cell Care Lab' is a growing company in the health-tech field.
Recently, an M&A with a major pharmaceutical company was decided, and the PR team was busy with 'announcement design'.

The lead, PR Manager Koji Ueda, created the press release materials and shared them internally.
However, an unexpected reaction came back.

• Internal: Slack went quiet, with few voices of celebration or expectation

• External: Inquiries from investors and business partners that sounded 'unsatisfied' appeared here and there

• Media: Only the headlines took the lead, reporting sensationally as the 'acquired side'

M&A: 'For whom and for what purpose is it?'
If you misjudge the center of gravity of the message, trust will collapse in an instant.
Ueda decided to rethink the communication design once more with AI.

■ Case: Changing the 'received words' for each stakeholder

First, he used Perplexity to research 'recent trends and response examples for M&A announcements'.
He identified the tone of message design and the factors that led to misunderstandings.

Next, he uploaded internal materials, past PR announcements, FAQ documents, and investor communications to NotebookLM to conduct consistency and tone analysis.

Prompt: 'Read the following internal and external messages and analyze where they might cause misunderstandings for each stakeholder. Organize and reconstruct the phrasing and priorities for internal staff, investors, media, and partner companies.'

As a result, it became clear that 'reassurance and future roles' for internal staff, 'business continuity and synergy effects' for external parties, and 'mid-to-long-term growth strategy' for investors were missing.

ChatGPT was used with the following prompt to create message drafts with words tailored to each reader.

Prompt: 'Regarding the M&A announcement, please create separate explanatory texts for 'internal staff, investors, and media'. Focus on the points of most concern for each position, and choose words that reduce the room for misunderstanding.'

Furthermore, Claude was used to refine the draft of the CEO's message into a "consistent, hopeful tone," and Canva was used to replace the presentation visuals.

■ Solution: "Addressing everyone" resonates with no one

Instead of relying solely on a press release, we prepared "briefing sessions," "QA documents," "CEO videos," and "Slack comments."

As a result—

• Employees: Anxiety about the future was eased, and a sense of security increased by 60% in employee surveys.

• Investors: They understood the mid-to-long-term M&A strategy, and inquiries shifted to those reflecting trust.

• Media: Articles with an accurate tone increased, and sensationalist coverage decreased.

Ueda realized something.
What should be questioned in communication design is not "what I want to convey," but rather **"who I want to reach and how I want them to receive it."**

■ Summary of AI Tools and Prompts Used

Tool | Usage Scenario | Prompt Used (Excerpt)
Perplexity | Researching past M&A announcement cases | "Please tell me about cases of M&A PR that led to backlash or confusion in recent years and the reasons behind them."
NotebookLM | Analyzing past documents and FAQs | "Please identify potential misunderstandings by stakeholder for this M&A document and suggest a restructuring plan."
ChatGPT | Paraphrasing messages | "Please generate separate messages for internal staff, external parties, and investors that convey the same intent."
Claude | Standardizing the tone of the CEO's message | "Please rewrite this M&A message in a positive and sincere tone."
Canva | Visual design for M&A explanatory materials | "Please propose a presentation template that conveys the background and future of the M&A."

■ Column: M&A PR is a job of becoming a "Translator of Trust"

M&A is not just about buying and selling.
It is a turning point that affects the "meaning of work," "trust," and "expectations" of everyone involved.

Speaking the same information in multiple languages.
That translation work is exactly what PR is.

AI can become an excellent "interpreter" to support that translation.
We must never forget "whose emotions we are protecting."

■ What to do in these situations?

Q: The media keeps emphasizing the "acquired side" in their reports.
→ Use ChatGPT to generate a draft that shifts the tone to "emphasizing the intent and future of the integration."

Q: Employees seem anxious about their "own future."
→ Use Claude to refine a message in a "CEO-to-employee tone that delivers reassurance."

Q: Investors are doubting the basis of the figures.
→ Use NotebookLM to verify consistency with the mid-term management plan, and use ChatGPT to create supplementary Q&A.

■ Reader Workshop: Try creating "drafts by audience" for an "M&A explanation"

1. Have ChatGPT generate "explanatory text for each stakeholder"

2. Refining the tone of comments from the CEO and management using Claude

3. Polishing "explanatory materials that go beyond just numbers" with Canva

4. Checking for contradictions and omissions against existing documents with NotebookLM

■ In Conclusion

While an M&A event may look like a "strategy" from the outside, on the inside, it is pure "emotion."

Even a slight difference in the words used to convey it can turn it into either anxiety or hope.

Public relations is "the work of translating information and connecting trust."

Generative AI helps untangle complex webs of words and provides the power to tailor them to the perspective of the audience.
With that power at your disposal, let's create "communications that support your company's future."

76 | Industry-Academia Collaboration PR: Designing a Human Touch That Goes Beyond "Just Results"

(Chapter 8 | Fulfilling Accountability to Systems, Management, and Society)

■ Introduction: "Why this university?"

Medical device manufacturer "Meditech Innovation" announced a new industry-academia collaboration project with one of the country's leading engineering universities.
The joint research and development of an AI diagnostic support tool was supposed to begin—or so they thought.

PR representative Natsuko Kiriyama prepared a press release and a landing page, and distributed them both inside and outside the company.
However, the response was lackluster.

• "It just looks like a typical joint research project."

• "Why this university? What makes it different from others?"

• "It sounds difficult. Who is it actually for?"

So, Kiriyama decided to change her perspective and redesign it starting from the "human touch."
She decided to have generative AI help her think through the PR perspective.

■ Case: Turning the "Background" and "Motivation" of a Project into a Story

First, Kiriyama used Perplexity to collect examples of "successful industry-academia collaborations" and "memorable PR expressions."
The keywords were "people," "passion," and "starting from the problem."

Next, upload the research summary, professor interviews, internal reports, etc., to NotebookLM and enter the following prompt.

Prompt: "Based on the following information, summarize the 'origin story' of this industry-academia collaboration project. Please include not only the technical elements but also the sense of purpose, the researchers' aspirations, and the social background."

What the AI returned was the fact that "the project began when the founder's daughter was diagnosed with a rare disease."
I was convinced that this 'starting point' was the story that needed to be told.

Furthermore, use the following prompt in ChatGPT to generate a structure where 'social issues' and 'personal aspirations' intersect.

Prompt: "Please restructure the following joint research project as a 'PR story.' Create a structure where social issues, technological innovation, and the researchers' aspirations overlap. Use an empathy-based context that makes the reader 'want to cooperate.'"

I polished that draft into a natural tone using Claude and prepared visual materials for the background and story in Canva.

■ Solution: What should be conveyed is not 'results' but 'ambition'

Due to the redesign of the PR, the response after dissemination changed significantly.

• Voices of 'empathy' arrived from students and researchers, which also spread to recruitment events

• It was favorably covered by the media as 'industry-academia collaboration where the passion is visible'

• Within the company, there was an increase in reactions from employees outside the technical department, saying 'our company is great'

Kiriyama realized once again the power of talking about 'why we are taking on this challenge now,' rather than numbers or technology.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt) | Perplexity | Researching success stories and collecting references for story structure | "Please tell me about PR examples of industry-academia collaboration that inspired people." | NotebookLM | Organizing background information and extracting 'story elements' | "Please extract points from the research background that could become the core of a 'PR story'." | ChatGPT | Story structure incorporating social issues and technology | "Please restructure the significance of the joint research in the context of 'social contribution and challenge'." | Claude | Polishing the tone with an emotional narrative | "Please add softness and enthusiasm to this text to make it reach people's hearts." | Canva | Creating story materials and visuals for dissemination | "Please suggest a story-driven material template that fits this research background."

■ Column: What PR needs is not information, but 'story design'

The goal of industry-academia collaboration is not just to 'announce it'.
What PR should convey is 'why this collaboration' and 'who is tackling what kind of problem'.

By showing the 'emotional motivation' before the results or technical jargon, people will listen.

Generative AI can verbalize that motivation, contextualize it, and transform it into the 'narrative' necessary for PR.

■ What should I do in these situations?

Q: The collaboration content is too technical and difficult to explain
→ Let's have Claude convert it into an 'easy-to-understand explanation for non-experts'.

Q: The reasoning for choosing this university is not very persuasive.
→ It is effective to organize researcher statements and backgrounds using NotebookLM, and then summarize the "shared vision" with ChatGPT.

Q: I want to explain this internally as well, but there is a gap in understanding.
→ Designing "internal newsletter slides" in Canva and reconstructing them in a narrative style will help get the message across.

■ Reader Workshop: "Discovering your company's 'Challenge Story'"

1. Upload project-related materials to NotebookLM

2. Execute a prompt in ChatGPT to "extract the 'human background' of this initiative"

3. Refine the tone with Claude and rewrite it as a story

4. Visualize the diagrams and introductory story in Canva

■ Conclusion

People are not moved by numbers or systems.
What moves them are "human passion," "doing something for someone," and "taking on challenges."

Industry-academia collaboration PR is not abouttranslating informationbut aboutdesigning a narrative.

Generative AI is "another editor" for putting that story into words.
Let's carefully tell the story of your company's "challenge."

77 | How to maximize the "power of words" through collaboration between lobbying, PA activities, and PR

(Chapter 8 | Fulfilling accountability to systems, management, and society)

■ Introduction: The disconnect between "policy advocacy" and "PR"

Environmental startup "Green Nest" is currently developing regional revitalization projects utilizing decarbonization technology.
As they deepened cooperation with national and local governments, their lobbying (policy advocacy) and public affairs (PA) activities also began in earnest.

However, PR representativeTakayuki Iguchifelt a sense of unease.

• "The advocacy materials are full of jargon, and I don't feel like they are getting through."

• "We aren't gaining empathy for why this system needs to be changed now."

• "We aren't fully communicating the social significance of our activities inside or outside the company."

With this in mind, Iguchi set out to use generative AI to translate 'specialized knowledge' into 'social impact'.

■ Case: Replacing 'Policy Proposals' with 'Someone's Future'

First, Iguchi uploaded the proposal documents and past lobbying memos to NotebookLM. Using the following prompt, he extracted elements that could convert specialized information into a reader-friendly perspective.

Prompt: 'Based on the following policy proposal documents and background information, extract keywords and structural elements that will resonate from a citizen's perspective. Instead of explaining complex systems, focus on "who it is for and how things will change."'

The extracted elements included angles connected to daily life, such as 'increasing the local employment rate for young people' and 'energy self-sufficiency during disasters.'

Based on this, he created a story-driven structure using ChatGPT.

Prompt: 'Reconstruct the following policy reform proposal into a story featuring a "family living in a rural area" as the protagonist. Create a structure that conveys the necessity of the policy or system as a story about protecting someone's future.'

Furthermore, he refined the tone to be more readable using Claude, and visualized the story materials with Canva. Finally, he researched similar success stories with
Perplexity to increase persuasiveness.

■ Solution: Designing 'Empathy Language' to Move Society

Iguchi's 'translation-style PR' was well-received by government agencies, industry associations, and local media.

• Government side: 'This could be used for citizen briefings. It's very helpful.'

• Internal team: 'We were able to realize how our work impacts society.'

• Local newspaper: 'It combines expertise with a human touch, making it easy to write an article about.'

Iguchi realized that systems and proposals become words that can move people when they are retold as 'someone's future'.

■ Summary of AI Tools and Prompts Used

Tool Usage Scenario Used Prompt (Excerpt) NotebookLM Organizing elements of policy documents and extracting 'translatable points' 'Extract keywords to transform specialized documents into a structure that resonates from a citizen's perspective.' ChatGPT Storytelling and structure creation 'Reconstruct the policy reform as a story about a certain family.' Claude Tone adjustment and readability optimization 'Refine the policy document into soft, easy-to-read text.' Canva Visual material creation 'Illustrate the background of this policy as a story about protecting the future.' Perplexity Researching success stories 'Tell me about policy proposal and lobbying cases that gained empathy from a citizen's perspective.'

■ Column: Policies Reach People as 'Someone's Story'

Systems are abstract. Policies feel distant. That is precisely why we need a perspective that
draws them closer to the 'lives of the people' behind them.

Public relations has the power to translate these things. Generative AI is a partner that will help you with that 'translation.'

'For whom, why now, and what kind of future are you creating?' Your words might be the key to moving policy.

■ What to do in these situations?

Q: The lobbying team and the PR team have different perspectives.
→ Use NotebookLM to organize materials from both sides, and have ChatGPT articulate a "common goal."

Q: Documents for government agencies are too stiff.
→ Instruct Claude to "rewrite in simple language" to change the impression.

Q: PA activity details are not being communicated internally.
→ Use Canva to design an "internal newsletter slide" centered on diagrams, and reconstruct it in a narrative style.

■ Reader Workshop: Translating "Institutional Language" into "Future Life"

1. Upload policy proposals and government documents into NotebookLM

2. Execute a "reconstruct from a consumer perspective" prompt in ChatGPT

3. Refine the text to be more accessible using Claude

4. Design a "story diagram" in Canva and deploy it for internal and media use

■ Conclusion

Lobbying and policy proposals are not merely "imposing systems," but
proposals to the question of "how to create a better society."

PR is the entity capable of translating those questions into the language of daily life.
And generative AI can become an "empathy designer" in that translation process.

Don't just convey; deliver.
One step beyond that lies the language that moves society.

78 | The Art of Re-editing IR and ESG Consistency as a Corporate Story

(Chapter 8 | Fulfilling Accountability to Systems, Management, and Society)

■ Introduction: "Is it okay to create IR and ESG separately?"

TechnoLinks, a global manufacturing company.
They produce both IR materials for investors and ESG reports conveying their approach to social issues every year.
However, the PR manager,Yoko Miyake felt a sense of unease.

• "IR emphasizes only 'sales and profits,' while ESG emphasizes only 'social contribution'"

• "When investors look at ESG, what should they actually believe?"

• "Even within the company, there's an atmosphere that 'ESG is the CSR department's job'"

Miyake set out to design materials that connect IR and ESG as a single corporate story.

■ Case: Aligning the “Corporate Voice” through “IR x ESG Cross-Editing”

First, Miyake uploaded the following materials to NotebookLM all at once.

• IR presentation materials from the past 3 years

• ESG reports

• Corporate philosophy and medium-term management plan materials

He used the following prompt to identify message consistency and points of contradiction.

Prompt: “Compare the following IR materials, ESG reports, and corporate philosophy materials, and analyze whether there is ‘story consistency.’ In particular, check if the axis of ‘what the company aims to be’ is wavering, and provide suggestions for improvement.”

Next, he used ChatGPT to create an outline that connects the ‘IR’ and ‘ESG’ materials with a ‘common storyline.’

Prompt: “Reconstruct the opening messages of the IR presentation and the ESG report, centered on the company’s ‘sustainable growth story.’ Please use a tone where ‘profit growth’ and ‘social value creation’ are naturally connected.”

Furthermore, he used Claude to refine the CEO’s message into a tone that resonates with both investors and consumers, and created an ‘IR x ESG integrated visual’ using Canva.

■ Solution: Design capability to connect the ‘story of numbers’ and the ‘story of society’

Miyake reaffirmed his role in PR as a ‘story editor.’

• The ‘growth strategy’ told in IR and the ‘social contribution’ told in ESG are actually the same ‘corporate vision’

• Connecting ‘Why grow this business?’ and ‘What kind of social value does it create?’ with a single line

• Making materials that allow the reader to easily understand ‘where this company is heading’

As a result, feedback from investors, employees, and business partners increased, with comments like ‘the tone of the materials has become easier to understand’ and ‘the company’s vision has become clear.’

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Used Prompt (Excerpt) | NotebookLM | Consistency and contradiction check between materials | ‘Compare IR and ESG materials, analyze if the axis of the ‘corporate vision’ is wavering, and provide improvement suggestions’ | ChatGPT | Story outline and structure creation | ‘Create a presentation structure for a ‘corporate growth story’ that connects profit growth and social contribution’ | Claude | Tone adjustment and unification of CEO message | ‘Create a CEO message that reaches both IR investors and general consumers’ | Canva | IR x ESG integrated visual creation | ‘Illustrate a corporate story that connects business growth and social value’

■ Column: "IR materials are about numbers, ESG is about passion"... Is that really enough?

"IR is logic, ESG is emotion"—
Are you creating materials with such a divide?

Investors and society alike are now looking at the "story" of what kind of future this company is aiming for.

Both the numbers and the initiatives are all "foreshadowing for the future."

AI can be a powerful assistant for weaving disparate documents into a single narrative.

■ What to do in these situations?

Q: The tone of my IR materials and ESG report is inconsistent
→ Have NotebookLM read both and ask it to extract a "common story axis."

Q: My IR materials tend to be just "explanations of numbers"
→ Ask ChatGPT to "tell this IR content as a business growth story."

Q: The CEO's comments sound like a different person in ESG vs. IR
→ Have Claude rewrite them in a "consistent tone for both."

■ Reader Workshop: Designing an "IR x ESG Storyline" for your company

1. Use NotebookLM to read past IR/ESG materials and extract the core themes

2. Use ChatGPT to create an "integrated story outline"

3. Use Claude to adjust the CEO's comments to a consistent tone

4. Use Canva to illustrate the "connection between business growth and social contribution" and share it internally and externally

■ Conclusion

IR and ESG are not two separate documents.
"They are two ways of telling a single corporate story."

That "connecting role" is the true essence of public relations.

AI can act as that adhesive.
Drawing the company's future in a single line—
That is the "editorial power of language" required of PR professionals from now on.

79 | Press conferences and presentations: What are the "scripts and lines" that don't leave the audience behind?

(Chapter 8 | Fulfilling accountability to systems, management, and society)

■ Introduction: "I have the materials. But what should I say and how?"

Nature Foods is a rapidly growing company in the food industry.

With the launch of a new sustainable brand, it was decided that the company would hold its first large-scale press conference. PR leader
Yusuke Sasaki was struggling with creating the script.

• "I've made the PowerPoint, but I don't see 'how to talk' along with these slides."

• "The CEO and directors are busy, so we can't take time for rehearsals."

• "It's dangerous to just 'wing it' during the Q&A session..."

The materials are there. But the design of "what to say, in what order, and how to tell it" is missing.

To make the press conference "resonate," Sasaki decided to use AI to reconstruct the "flow of words."

■ Case: Simulating "Dialogue" and "Emotional Flow" with AI

First, he uploaded the presentation slides for the day, past CEO speech drafts, and Q&A logs from previous press conferences to NotebookLM.

He analyzed the flow of the talk and common question patterns using the following prompt.

Prompt: "Based on the following press conference materials and past speech drafts, please rearrange them in a 'story order that is easy to understand.' Furthermore, analyze the past Q&A logs and create a list of anticipated questions and answers."

Next, he created a script draft to design the "listener's emotional journey" using ChatGPT.

Prompt: "Following this presentation material, design an emotional flow that makes reporters want to 'empathize, understand, and ask questions,' and output it as a 'spoken script' for the speaker to read. Please avoid difficult words and use a conversational tone."

Claude was asked to draft an "opening speech in a natural tone" for the CEO.

Furthermore, he created media handouts to be distributed on the day of the press conference using Canva.

■ Solution: Working backward from "how the listener receives it"

With the support of AI, Sasaki completed a "script that is easy to speak and easy to understand" in a short period of time.

• Talking about the "background and passion" at the beginning

• Emphasizing the "concreteness of products and initiatives" in the middle

• Structure the conclusion to speak about 'prospects for the future'

The Q&A session was prepared in advance with a list of anticipated questions.
On the day, it was well-received by reporters, who commented that it was 'easy to understand' and 'conveyed passion'.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Prompt Used (Excerpt)
NotebookLM | Analyzing the flow of slides and past drafts, and extracting Q&A | 'Reorder the materials in a way that conveys the message effectively, and create a list of anticipated questions.'
ChatGPT | Designing the script and flow of speech | 'Based on these materials, create a draft script in an order that moves the listener's emotions.'
Claude | Adjusting the tone of opening speeches and representative comments | 'Make this opening statement sound sincere and natural.'
Canva | Designing handout materials for distribution | 'Create a template for a one-page A4 handout that summarizes the content of this press conference.'

■ Column: PR scriptwriting is 'designing the listener's emotions'

A press conference is not just for 'conveying information,' but a 'place to create an atmosphere.'

Where will the listener be convinced, where will they become interested, and where will they want to ask questions?
It is the role of PR to weave that 'emotional path' into the script.

Generative AI is the ultimate 'dialogue partner' for simulating emotional flow and refining tone and wording.
If you are struggling with press conference preparation, first try having the AI imagine the 'listener's feelings.'

■ What to do in these situations?

Q: The president won't read the script...
→ Use Claude to create a 'short speech with a natural tone' and refine it into 'easy-to-speak language.'

Q: I'm afraid of the Q&A session and it takes too long to prepare
→ Feed past reporter question logs into NotebookLM and have it automatically generate an 'anticipated Q&A list.'

Q: I feel like I'm going to rely too much on the slides
→ Use ChatGPT to create a 'narration script' that allows you to speak without looking at the slides.

■ Reader Workshop: 'Press Conference Script Emotional Design Workshop'

1. Analyze past speech and Q&A data with NotebookLM

2. Create a draft script based on 'listener emotions' with ChatGPT

3. Adjust the president's comments with Claude

4. Finalize handout materials and visual aids for the press conference with Canva

■ Conclusion

A press conference is a place to convey information, but it is also a stage to generate empathy.

By designing an 'experience' rather than just 'information' together with AI,
you can realize a press conference that reaches the hearts of your listeners.

Once the materials are ready, the next step is to create the right "atmosphere."
Let's refine the dialogue and timing for that together with AI.

80 | 3 Time-Saving Measures and AI Utilization Methods to Use When You "Don't Have Time for PR"

(Chapter 8 | Fulfilling Accountability to Systems, Management, and Society)

■ Introduction: "Too many tasks, can't keep up"

Mid-sized IT company "Connect Solutions."
250 employees, with a PR team of only 2 people.

PR Manager Ayaka Murase was chased by this situation every day.

• "I haven't even started on the next press release..."

• "The internal newsletter draft and social media updates are also being pushed back."

• "Interview requests, internal coordination, approval flows... everything is happening at once."

At that moment, a colleague said to her:

With that one remark, Murase began to seriously consider "saving time in PR."

■ Case: From "Doing it all at once" to "Delegating to AI"

The first thing Murase tackled were the following "3 time-saving measures."

① Saving time on "drafting the outline" of press releases with AI

ChatGPT is used to input past releases and bulleted notes, creating a draft with the following prompt.

Prompt: "Based on the following factual information, please create a basic structure for a press release (title, lead paragraph, and 3 body paragraphs)."

With this, the time spent writing from scratch was cut in half.

② Leaving "multi-style conversion" for social media and intranets to AI

Use the same information with Claude to expand into multiple tones using the following prompt.

Prompt: "Based on the following press release content, please convert it into drafts for internal intranet, X, LinkedIn, Instagram, and email newsletters. Please use a tone suitable for each medium."

This drastically reduces the need for "copy-paste and edit" work.

③ Save time with "AI summary templates" of past cases

NotebookLM Upload the past year's worth of PR achievements, internal articles, and coverage lists.

Prompt: "From the following past PR cases, extract the 'frequently used expression patterns' and 'structures that received good responses,' and create a reusable template."

This creates a system where you can start working without hesitation when similar projects arise.

■ Solution: From "doing everything yourself" to "dividing tasks that should be left to AI"

Murase realized something.

• Humans handle planning and intent determination

• AI handles information organization, first drafts, and tone adjustment

• Humans handle final checks and polishing the language as the 'face of the company'

By making this distinction, monthly PR work time was reduced by 30%. Even so, the quality of the output actually improved.

■ Summary of AI tools used and prompts

Tool | Usage Scenario | Used Prompt (Excerpt) | ChatGPT | Press release outline creation | "From the following information, create a release draft with a title, lead paragraph, and three body paragraphs" | Claude | Draft conversion for multiple media | "Convert the style of this content for intranet, SNS, and email newsletters" | NotebookLM | Template generation from past cases | "Analyze past cases and summarize the 'high-impact structure templates' for PR documents"

■ Column: "'Doing all PR work yourself' is already outdated"

The value of PR is not in "how much manual work you did," but in "how much impact you made."

To achieve that, focus on "work that only humans can do." You can "completely delegate" information organization, first drafts, and tone adjustment to AI.

What is important is to create a system where "you can focus on the final polish."

■ What should I do in these situations?

Q: The text AI produces is mediocre, and it actually increases the amount of editing I have to do...
→ Try refining your prompt in ChatGPT by saying, "Use past company press releases as a reference to match the writing style."

Q: Adjusting the tone for each SNS is a hassle.
→ You can save time by giving a bulk instruction in Claude: "Optimize this same content for multiple social media platforms."

Q: I don't have the bandwidth to look back at past cases.
→ It is effective to have NotebookLM perform a "pattern analysis of the past year's PR activities" all at once.

■ Reader Workshop: "Try creating your own 'AI time-saving PR process'"

1. Standardize the press release draft creation flow with ChatGPT

2. Develop a tone conversion flow for multimedia using Claude

3. Build a "past case study template collection" with NotebookLM

4. Use Canva to visualize time-saving measures for internal sharing and explain them to relevant departments

■ Conclusion

"Not having enough time" is the eternal challenge of PR.
But the times have changed.

By using AI as your "alter ego," PR can evolve from "I can't do it because I don't have time" to "I can manage things smartly with AI precisely because I don't have time."

The first step is to decide what to delegate and what to do yourself.

Now, welcome AI as your "time-saving partner" into your PR work.

81 | I want to get on TV and in the Nikkei newspaper! How to find "story ideas" that work in the field

(Chapter 9 | Applications and Practices for Expanding the Power of PR to Society)

■ Introduction: "What kind of stories will you cover?"

Sky Data Solutions, an IT venture in its 10th year of business.
PR representative Shuichi Nakanishi had recently been struggling with a certain issue.

"We want to get on TV and in the Nikkei newspaper more"—this has been the management team's long-held wish.

However, even when he sent pitches to reporters, he got zero replies.
The only things that had been published in the past were in industry papers and web media.

• "Is our news really that dull?"

• "How can I make them think, 'I want to cover this'?"

• "What is the 'granularity of a story' that the Nikkei would bite on?"

Nakanishi decided to use AI to objectively analyze his company's "news value."

■ Case: Conducting a "News Value Diagnosis from a Reporter's Perspective" using AI

First, Nakanishi usedPerplexityto research "articles on IT companies covered by the Nikkei Shimbun and TV stations over the past year."

Next, he uploaded his company's past press releases, product announcement materials, columns, and recruitment PR achievements toNotebookLM.

He checked the "story value from a media perspective" using the following prompt.

Prompt: "Based on the following company content, extract topics that have the potential to be covered by the 'Nikkei Shimbun' and 'TV news.' Please output them in a ranking format based on news value from the perspectives of 'social relevance,' 'novelty,' and 'impact.'"

The results included candidates such as "industry-first data linkage technology," "social contribution projects supporting small and medium-sized enterprises," and "productivity improvement measures through human resource diversification."

Furthermore, he executed the following prompt inChatGPT.

Prompt: "For the following story candidates, create 3 patterns each of press release title drafts and lead paragraph drafts from the 'angles that reporters want to know.' Please provide them from the perspectives of social relevance, impact, and human interest stories, respectively."

Claudewas used to prepare a "talk script for reporters," and a "one-page visual document for reporters" was completed inCanva.

■ Solution: "Our 'commonplace' becomes 'novelty' for society"

What Nakanishi realized was that "stories that are too commonplace for us to notice" are fresh and socially relevant information when viewed from the outside.

• What the media wants is not 'facts,' but a 'story to present to society'

• Structure, angles, and questions that make it easy for reporters to write articles are important

• The process of selecting stories through the 'eyes of a reporter' rather than 'PR subjectivity' is essential

As a result, he achieved his first publication in the Nikkei Shimbun's web edition and was also covered by local TV stations.

■ Summary of AI tools and prompts used

Tool | Usage Scenario | Usage Prompt (Excerpt)
Perplexity | Researching past media coverage | "Please tell me about IT company cases featured in the Nikkei newspaper and on television over the past year."
NotebookLM | Diagnosing the news value of in-house content | "From the following materials, rank them in order of 'high news value' from the perspectives of social relevance, novelty, and impact."
ChatGPT | Creating press release titles and lead paragraphs | "Please provide title and lead paragraph drafts from the three perspectives of social relevance, impact, and human story."
Claude | Creating talk scripts for reporters | "Please create a 30-second talk script to explain this news to a reporter."
Canva | Creating one-page materials for reporters | "Please suggest a template for one-page materials for reporters that summarizes this story content in an easy-to-understand way."

■ Column: 'Company Norms' are 'Fresh Social News' when viewed from the outside

The perspective that PR professionals most often forget—
It is that 'our daily routine might be extraordinary to society.'

Whether it is 'technological evolution,' 'changes in work styles,' or 'community involvement,'
they can transform into 'social news' depending on how they are presented.

AI can highlight internal information as 'socially relevant stories.'
When PR professionals become too accustomed to the 'norm,' the objectivity of AI is effective.

■ What to do in these situations?

Q: I don't know which news has 'high media value'
→ Let's summarize past internal materials in NotebookLM and have ChatGPT generate a 'media-friendly story ranking'.

Q: My lead paragraphs are always dull and don't resonate with reporters
→ Let's have ChatGPT generate title drafts and lead paragraphs from the three perspectives of 'social relevance,' 'impact,' and 'human story'.

Q: I am not good at approaching reporters by phone
→ It will go smoothly if you try after generating a '30-second explanation script for reporters' with Claude.

■ Reader Workshop: 'Discovering In-house Stories & Creating a 3-Perspective Ranking'

1. Summarize past PR materials and business topics with NotebookLM

2. Generate story ideas with ChatGPT based on the three axes of 'social relevance,' 'impact,' and 'human story'

3. Prepare a 'talk script for reporters' with Claude

4. Visualize simple materials for distribution to reporters with Canva

■ Finally

The weapon of PR is the 'power of discovery.'

Before lamenting that 'there are no stories,'
let's work with AI to turn our company's 'norms' into 'social news.'

That shift in perspective will lead to the next piece of coverage.

82 | Social Issue Solving PR: How to turn 'correctness' into 'empathy'?

(Chapter 9 | Application and Practice of Spreading PR Power to Society)

■ Introduction: "Why doesn't it resonate even when I'm saying the right things?"

Green Leaf Solutions is a startup gaining attention in the sustainability sector.

They completed a new eco-friendly packaging product and issued a press release as a social issue-solving PR initiative. The PR representative,
Misaki Ishikawa felt confident—but the results were disastrous.

• "Zero inquiries"

• "Zero social media engagement"

• "Zero media coverage"

Within the company, there was a feeling of, "It's such a great initiative, so why isn't it spreading?"

Ishikawa thought anxiously.

What was needed next was a "design that generates empathy."

■ Case: The process of transforming from "correctness" to "story"

Ishikawa first uploaded the following information to NotebookLM.

• Press release draft

• Internal documents (product development background, target customer insights)

• Past social media post data

She then entered the following prompt.

Prompt: "Based on the following materials, point out where this PR is too focused on 'correctness.' If I were to add 'empathetic story elements' for the reader, where and what kind of angle would be effective? Please propose suggestions."

Next, she created a "story-based press release" using ChatGPT.

Prompt: "Please create a story-style press release about this new product from the perspective of a 'certain consumer's' daily life. Please explain 'why this product was necessary' starting from 'everyday struggles.'"

Furthermore, she refined the tone to be more natural using Claude, and also created "empathy-driven social media post visuals" using Canva.

Finally, I used Perplexity to research "buzz-worthy social issue resolution PR cases" and used it to help tune the title and structure.

■ Solution: Conveying "someone's feelings" rather than "issues"

The answer Ishikawa arrived at was simple.

• "Social issues" are "situational explanations"

• "Empathy stories" are "emotional hooks"

"A housewife's shopping scene," "a child's remark," "a developer's setbacks and challenges"—by changing it into a story that carries such "human emotions," it spread on social media immediately after the release was distributed.

It was also picked up by the media as a "product with a human drama."

■ Summary of AI tools used and prompts

Tool | Use Case | Prompt Used (Excerpt) | NotebookLM | Analysis of "excessive correctness" in current drafts and improvement suggestions | "Point out parts of this PR document that are 'too focused on correctness' and suggest ways to turn them into an 'empathy story'" | ChatGPT | Creation of story-style releases | "Please create an empathy-style release for this new product from the perspective of 'a consumer'" | Claude | Tone adjustment and readability improvement | "Please refine this draft into a 'natural and soft tone'" | Canva | Creation of visuals for social media posts | "Please design an empathy-based social media post visual that matches this product and story" | Perplexity | Research on successful social issue resolution PR cases from other companies | "Please tell me about domestic and international social issue resolution PR cases that evoked empathy"

■ Column: Social issues are not "announcements of the correct answer" but "stories of empathy"

"I'm saying the right things, but it's not resonating"—the reason is that there is no "hook for the reader's heart."

People are moved by human thoughts, not data.

That is why what PR should do is not "presentation of facts" but "translation of emotions." AI supports that "translation work" quickly and accurately.

■ What should I do in these situations?

Q: I'm afraid to cut facts, so my releases become information-heavy
→ Use NotebookLM to perform a "sufficiency check," and have ChatGPT adjust it to the "optimal volume from the reader's perspective."

Q: I'm worried that making it emotional will make it seem fake
→ Having Claude create a "fact-based, empathy-evoking narrative" will result in a natural tone.

Q: I want short, emotionally-hooking text for social media
→ Have ChatGPT output multiple "empathy-based caption drafts for social media."

■ Reader Workshop: "Correctness to Empathy Story" Conversion Workshop

1. Load your company's social issue-related PR materials into NotebookLM and check for sufficiency

2. Create multiple patterns of "consumer perspective story drafts" using ChatGPT

3. Refining the natural tone of voice with Claude

4. Creating SNS visuals that lean into 'emotional hooks' with Canva and conducting internal reviews before distribution

■ In Conclusion

What resonates in social issue PR is 'emotion' rather than 'facts'.

People are not moved by 'correctness' alone.
What is needed is the PR design capability to deliver it as 'someone's story'.

If you refine that process together with AI,
your words will surely reach society.

83 | How to overcome the 'only the president can speak' problem in SME public relations

(Chapter 9 | Applications and Practices for Expanding the Power of PR to Society)

■ Introduction: 'Always relying on the president's comments... is this okay?'

Takashima Tech, a company involved in the manufacturing and sales of construction-related equipment.
It has about 70 employees and is a long-established company that has been operating in a regional city for over 40 years.

PR representative Tsuyoshi Taniguchi had been worried about this for a long time.

• 'Media interviews = the president always appears'

• 'Event appearances, contributed articles, all the president'

• 'But the president is busy, and there's a limit to generating ideas...'