[Event Report] "AI Writing Support Review Plus" Service Briefing Held on 8/22

Realizing efficiency in article creation “AI Writing Support Review Plus”. How was this service, which automatically generates first drafts from audio/text/video data and even checks for risks of public backlash, born?

Tsukuma from note inc. and Tanaka, the development lead at note AI creative inc., spoke in detail about the background of the development and how to use the service.

*This article was written using "AI Writing Support Review Plus" and was published after editing.

Overview of "AI Writing Support Review Plus"

──First, could you tell us what kind of service AI Writing Support Review Plus is?

Tsukuma　AI Writing Support Review Plus is a tool that creates first drafts of interview articles from data such as audio, text, and video. In addition to creating articles, it is also equipped with a feature to check for risks of public backlash. We began general sales on July 31, 2025, and released it for corporations and sole proprietors.

──I see, so it even checks for risks of public backlash. Could you tell us a bit more about the features?

Tanaka　This is a service that generates a first draft using AI, and it has the following three review functions.

• Text review

• Character count adjustment

• Review for risks of public backlash

With these features, it automatically suggests revisions and adjusts the manuscript to be more readable. Additionally, the first draft is provided in two versions: a long-form version and a slim version with a condensed character count.

──That is a wonderful feature! What are the actual benefits of implementing it?

Tanaka　By implementing this service, you can significantly reduce the time it takes to create interview articles.

Also, since the delivered items include the full transcript, there is no need to perform the traditional transcription work after an interview. Another benefit is that you can focus more on the interview itself.

Development Background and Philosophy

──To begin with, what was the trigger for developing this service?

Tsukuma　It started with an in-house project to improve operational efficiency using AI.

In particular, we found that the burden of writing and editing interview articles for in-house editors was significant. Creating interview articles takes a lot of time, including preliminary research, question preparation, conducting the interview, editing, and communication with stakeholders.

It started from the thought of whether we could develop something to support those tasks.

──That certainly sounds like a difficult task. Could you also tell us about the development philosophy?

Tanaka We aimed to enable article creation using the same workflow as an editor. To that end, we designed it so that editors and writers can work easily and achieve a level of quality they are satisfied with.

What I was particularly conscious of was not stripping away too much information in the first draft. AI is good at summarizing long texts, but details can be lost in that process. For an editor, the psychological burden of trimming down a manuscript with a lot of information is lower, whereas the task of searching for and supplementing missing information is a significant cost.

That is why we prioritized creating a manuscript that retains as much information as possible in the first draft. I believe this allows editors to improve quality while cutting out unnecessary parts.

▼ For more details, please see the following article.

Effects of Introduction

──In reality, what kind of effects have you seen since introducing the service? Could you give us some concrete numbers?

Tsukuma At our company, it used to take about 36 hours to create an interview article before introducing AI. This included everything from preliminary research and accompanying interviews to communication with partner companies. On the other hand, when using AI Writing Support, we were able to create it in 8 hours.

──From 36 hours to 8! That is a huge reduction. How about the usage track record?

Tanaka The AI Writing Support provided for note pro, which is the predecessor to AI Writing Support Review Plus, is used by over 140 companies, and the introduction cases for Calbee, Inc. and the Kyoto City Government have been made public.

Calbee evaluated us for our high accuracy and time-saving effects, and the Kyoto City Government told us that the peace of mind provided by AI allowed them to focus on what the interviewee was saying. I believe this is a wonderful example of utilization that achieves operations based on the premise of AI.

──That is a wonderful track record. Can this service be used for things other than interview articles?

Tsukuma Of course. You can significantly shorten the time it takes to create various content that conveys raw voices, such as customer usage cases and messages from the president, in addition to employee interviews. Since it can also be used for articles published outside of note, such as corporate websites and recruitment sites, I hope those who have audio or video content will try using it.

How to Use

──What should I do if I want to use it?

Tanaka First, please apply for a trial from the URL listed in our press release article.

Once you enter the required information in the application form and submit it, you will receive an email with instructions on how to start your trial. By creating an account from the link in the email, you can use AI Writing Support Review Plus for free for one month, once per company.

──What does the screen look like after creating an account?

Tanaka When you create an account, the top page will be displayed. If you select AI Writing Support Review Plus, you will be taken to the submission form below. On the submission form, you can upload audio or video and enter information such as the number of speakers to complete the task.

──Please tell us about the flow after submission as well.

Tanaka After submission, a guidance pop-up will appear. If you wait for a while, you will receive an email notifying you that delivery is complete, and you can check the contents by downloading the file from there.

──What is included in the delivered items?

Tanaka There are a total of six files in the delivered items. The first one is what we call the "first draft," which is a manuscript written by the AI based on the content of the interview. The results of the text review and the reputational risk review have already been reflected, making it a comprehensive manuscript with a large volume of text.

──I'm also curious about the "Slim Version." What kind of version is that?

Tanaka The "Slim Version" is a version with a more compressed word count than the first draft. We have shortened the text while ensuring that important points and the highlights of the interview are not left out of the manuscript. It can be used when you want to publish quickly or reduce revision costs.

──How can I check the results of the controversy review?

Tanaka The controversy review results are provided in a file that indicates the reasons for revisions and how to make them, in case there are any expressions in the interview that could potentially cause issues. The first draft and the Slim Version have already been created by reflecting these results, but by looking at this file, you can specifically understand the perspective from which the revisions were made.

──Please tell me the role of the outline file.

Tanaka The outline is a file that allows you to get an overview of the interview content and check the structure of the first draft. It is useful for grasping the overall content before reading the manuscript.

──How should I use the Q&A format file?

Tanaka The Q&A format file allows you to get an overview of the exchanges during the interview. When you want to check what kind of exchanges took place, looking at this file is helpful for understanding the content.

──What kind of information is included in the transcription file?

Tanaka The transcription file contains text with speaker separation and timestamps. Since you can see what kind of exchanges happened at what time, it is also convenient when checking the audio file. It also includes analysis results at the beginning indicating the roles of the speakers, so you can read the content after confirming who the interviewer is.

──So, the self-introduction and ice-breaking at the beginning are not included in the article?

Tanaka Since the AI determines what content should be included in the article, parts that are not the main topic, such as self-introductions, are not included. However, if you speak your self-introduction near the beginning of the interview, it makes it easier to identify the speakers when checking the transcription, so it might be good to keep that in mind.

Pricing Plans

──Please tell me about the pricing structure I'm curious about.

Tsukuma There are two types of pricing plans. The first is the "2 Articles Per Month Plan," which is free for the first month and 10,000 yen per month from the second month onwards. The second is the "Unlimited Monthly Plan," which is 80,000 yen per month.

──Is it possible to change plans?

Tsukuma Yes, it is possible to change plans. Also, even if you initially choose the 2-article plan, if you decide you want to create a 3rd article or more along the way, you can switch to the unlimited plan. However, if you use the 3rd article or more while on the monthly 2-article plan, you will need to sign up for the unlimited plan, and an additional fee will be incurred even in the first month. Conversely, it is also possible to switch from the unlimited plan to the 2-article plan.

Points on Data Security and Accuracy Improvement

──I am concerned about security. How is data safety handled?

Tanaka The data you upload is never used for AI training. The text output during the generation process is also not used for model training. Additionally, uploaded audio and video files are deleted from our system 10 days after registration, so you can use the service with peace of mind.

──By the way, are there any tips for improving the accuracy of the first draft?

Tanaka I believe the way you conduct the interview and record the audio is important. In addition to introducing yourself at the beginning, when you want to ensure certain points are included in the article, if the interviewer emphasizes them verbally by saying things like "That's interesting" or "That's a key point," the AI will pick up on those cues. It is important to leave verbal markers for points that are significant to both humans and the AI.

Differences from Generative AI and Writing Quality

──Please tell us the difference between this and general generative AI services.

Tanaka To create interview articles from audio or video using general generative AI services, you need specialized knowledge and skills such as prompt engineering. On the other hand, I believe the major difference with this product is that even without such specialized knowledge, you can simply upload audio or video to easily obtain high-quality manuscripts that include text refinement and reviews for potential controversy risks.

──I've heard that using generative AI can lead to unnatural writing. What are your thoughts on that?

Tanaka Regarding writing quality, this is a point we were careful about during service design. We have incorporated a polishing step to maintain natural flow between sentences, ensuring the text is refined into natural prose. Our in-house editors have also reviewed the manuscripts, and we believe the level of writing is high and that the so-called "AI-like" feel has been significantly reduced.

If you are interested in AI Writing Support Review Plus

▼Click here to apply

※This service is only available to businesses that are registered as corporations or have submitted a notification of opening a business as a sole proprietor.

For more details about "AI Writing Support Review Plus," please see this article.

※This article was written using "AI Writing Support Review Plus" and was published after editing.