The Complete Guide to Mastering Microsoft 365 Copilot

The AI assistant "Copilot," integrated into Microsoft 365, is fundamentally changing the way we work. Built into everyday applications like Word, Excel, and PowerPoint, this feature supports all kinds of tasks, from document creation and data analysis to presentation building. In this article, we will explain how to use Copilot in each application from a practical perspective, including points to keep in mind when implementing it.

Easy-to-understand audio delivery on Spotify: "Rami's AI Tech Radio"

According to a World Economic Forum survey, "curiosity and lifelong learning" ranked 5th among the "skills needed in 2030." Curiosity is an economic necessity, not a hobby. I explain this in detail in my book, "The Ultimate Skill for the AI Era: Curiosity Power."

Preface

Microsoft Copilot is not a single product, but a group of AI assistants designed to meet diverse needs, from improving individual productivity to transforming corporate operations. In this article, we will cover each edition that makes up the Copilot family (free version, Copilot Pro, Microsoft 365 Copilot, Copilot in Windows, Copilot in Edge, GitHub Copilot, Copilot for specific tasks, Copilot Studio, etc.) and explain their main features, target users, pricing, expected benefits of implementation, potential risks, and ethical considerations.

1. While Microsoft Copilot is based on OpenAI's LLMs, it stands out, particularly in Microsoft 365 Copilot, for its ability to provide highly personalized support by leveraging organizational data (emails, chats, documents, etc.) through integration with Microsoft Graph.

2. While Copilot holds great potential for significantly improving work efficiency and supporting creativity, it also entails challenges such as the accuracy of generated information, data privacy, and the protection of intellectual property rights. Thirdly, to maximize Copilot's capabilities, it is essential to select the appropriate plan for your usage purpose, develop effective prompt (instruction) creation skills, and engage in continuous learning as AI evolves.

Microsoft Copilot: AI-Powered Productivity Assistant

Core Concepts and Underlying Technology

Microsoft Copilot is an AI-powered digital assistant designed to understand user questions and requests and support a wide range of tasks, including providing information, assisting with writing, and creating images.

At its core is an LLM developed by OpenAI, which enables high-precision responses and content generation in a way that feels like a natural conversation with a human. LLMs are AI algorithms that use deep learning technology and vast amounts of text data to perform advanced processing such as language understanding, summarization, prediction, and the generation of new content.

In Microsoft 365 Copilot for business, in particular, a major feature is that the capabilities of this LLM are closely integrated with Microsoft Graph. Microsoft Graph is a gateway that enables access to Microsoft 365 data within an organization (emails, calendars, chats, documents, contacts, etc.), through which Copilot deeply understands the user's specific work context.

While many AI assistants are limited to general information and task processing, Microsoft 365 Copilot can generate personalized responses and suggestions "based on the user's own data." For example, it can draft replies based on past email exchanges or summarize reports based on internal documents.

This integration with organizational data is the key to elevating Copilot from a mere information search or text generation tool to a true "copilot" that is indispensable for daily work. This approach not only realizes more relevant and practical AI support for users but also has strategic significance in increasing the value of the entire Microsoft 365 ecosystem.

Overview of the Microsoft Copilot Family

Microsoft Copilot does not refer to a single product but consists of a very wide range of product groups, from free versions to high-functionality versions for corporations, specialized tools for developers, and even solutions specialized for specific business areas. These product groups are services that were previously provided under names like "Bing Chat" and "Bing Chat Enterprise," which were integrated and reorganized under the "Copilot" brand in November 2023.

This move toward brand unification can be interpreted as a clear strategy by Microsoft to permeate its AI assistant functionality throughout its entire ecosystem and provide users with a consistent user experience.

The fact that the name "Copilot" is attached to various Microsoft products and services, such as Copilot in Windows, Copilot for Sales, and GitHub Copilot, also shows the company's strong will to make AI available at every touchpoint.

As a result, no matter which Microsoft product users are using, they can receive AI support through the familiar "Copilot" interface, which is expected to reduce learning costs and promote usage.

Each edition of the Copilot family has its own characteristics in terms of target user base, provided features, pricing, and delivery format. Therefore, when individuals or organizations consider implementing Copilot, it is extremely important to carefully select the edition that best matches their needs and objectives.

If this choice is made incorrectly, you may not achieve the expected results or may incur unnecessary costs, so a detailed understanding of each edition is required.

Microsoft Copilot Editions and Service Details

Microsoft Copilot offers multiple editions and related services to meet diverse user needs. Understanding the features of each and selecting the most appropriate one is the first step toward effective utilization.

Microsoft Copilot (Free Web and Mobile Version)

Previously known as "Bing Chat," this service is the basic entry point for the Copilot family. Anyone can start using it for free via a web browser or the dedicated mobile app, regardless of whether they are using a PC, Mac, or smartphone. The primary target audience is general individual users who intend to use it for information retrieval, simple brainstorming, daily writing tasks, and image generation.

Key features include chat-based Q&A using LLMs, internet information retrieval, generation of text and images based on instructions, and summarization of long-form content. While you can use high-performance AI models, there are some limitations to the free version.

In particular, during times when there are many users, such as when servers are congested, paid users are given priority access, which may result in slower response times or temporary usage restrictions for free users.

An important point for business use is that signing in with a work or school Microsoft account (Microsoft Entra ID) enables commercial data protection features.

This alleviates concerns often associated with using free AI tools regarding the handling of input information and corporate data. This feature lowers the psychological hurdle for corporate users trying out the free version, providing an environment where they can use it with a certain level of peace of mind, even for prompts containing business-related information.

Microsoft providing such enterprise-level data protection even in the free version can be seen as a strategic move to gain the trust of corporate users and encourage future migration to paid versions.

Microsoft Copilot Pro (Paid Version for Individuals)

Microsoft Copilot Pro is a paid subscription service designed for individual users who want to further enhance the performance of the free version of Copilot and seek more advanced features and a more comfortable user experience.

The primary target audience includes users who want priority access to the latest AI models at all times, and those who want to maximize their personal productivity by utilizing integration features with Microsoft 365 applications used daily, such as Word, Excel, and PowerPoint.

The main benefits of Copilot Pro are as follows:

Priority access to the latest models
Priority access to the latest and most high-performance large language models is guaranteed, even during peak times. This makes it less susceptible to the response delays and usage restrictions sometimes seen in the free version, allowing for consistently smooth AI interaction.

Integration with Microsoft 365 apps
Users with a Microsoft 365 Personal or Family subscription can use Copilot features within major desktop applications such as Word, Excel, PowerPoint, Outlook, and OneNote (Windows version). This allows AI to directly support daily tasks such as document creation, data analysis, presentation material preparation, and email management.

Enhanced AI image generation
You can use faster processing (boosts) for AI image generation using the Designer feature (formerly Bing Image Creator) up to 100 times per day. This is a significant increase compared to the 15 times per day for the free version, which is a major advantage for users who frequently perform creative work.

The price is 3,200 yen per user per month, or 20 US dollars per month. Please note that a separate Microsoft 365 Personal or Family subscription may be required for use.

Microsoft 365 Copilot (Paid Version for Businesses)

Microsoft 365 Copilot is the most comprehensive AI assistant service for businesses, designed to improve corporate productivity and transform business operations. It is deeply integrated into the Microsoft 365 application suite that employees use daily, such as Word, Excel, PowerPoint, Outlook, and Teams, and functions not just as an add-on, but as a way for AI to support business processes themselves.

The most significant feature of this edition is its integration with the aforementioned Microsoft Graph. Copilot understands the vast amount of data within an organization—such as email exchanges, shared documents, meeting records, and chat history—and uses this to provide highly relevant assistance tailored to each user's current work context.

As a result, for specific instructions like 'Summarize the key decisions and action items from last week's Teams meeting regarding Project A,' Copilot can cross-reference and analyze relevant information to generate an accurate response.

The target users are organizations that already hold corporate licenses such as Microsoft 365 Business Standard/Business Premium, Microsoft 365 E3/E5, F1/F3, or licenses like Office 365 E1/E3/E5/F3. Dedicated licenses for educational institutions are also available.

While the main features are diverse, the core focus is on content creation support across all Microsoft 365 applications (e.g., drafting reports in Word), data summarization and analysis (e.g., analyzing sales data in Excel), task automation (e.g., creating routine emails in Outlook), and communication support (e.g., summarizing meetings in Teams).

All of these features are built on the principles of enterprise-grade security, privacy protection, compliance, and responsible AI, ensuring a design that companies can adopt with peace of mind.

1. Microsoft 365 Copilot Chat (formerly known as Microsoft 365 Chat)

One of the core features of Microsoft 365 Copilot is Microsoft 365 Copilot Chat (formerly Microsoft 365 Chat). This is an intelligent chat service that understands information across a user's entire Microsoft 365 environment—including chats, emails, files, meeting records, and calendars—and provides various forms of assistance in a conversational format.

The target users are Microsoft 365 subscription users with a Microsoft Entra ID account, and basic chat functionality is available at no additional cost. However, some usage-based fees may apply for the use of more advanced features or agent capabilities.

In terms of functionality, it is equipped with the latest LLM and applies enterprise data protection as standard. Furthermore, it offers diverse features such as the use of custom agents created in Copilot Studio, information provision through file uploads, and AI-powered image generation.

Users can seamlessly access Copilot Chat from within the Microsoft 365 applications they use daily, such as Teams, Outlook, and Microsoft Edge, to ask work-related questions, search for information, create content, and catch up on missed information.

Copilot in Windows (Windows OS Integrated Version)

Copilot in Windows is an AI assistant feature integrated directly into the Windows 11 operating system. By simply clicking the Copilot icon displayed on the taskbar, users can receive AI assistance at the OS level without switching applications. This feature aims to provide answers and inspiration from the web, support daily creative tasks, and help users stay focused on their current work.

The primary target users are all individuals using Windows 11. It is provided as a standard OS feature of Windows 11, and its availability is gradually expanding through OS updates, such as being introduced as a preview version starting from version 23H2.

Specific features include support for adjusting various PC settings (e.g., switching to dark mode, changing notification settings), organizing multiple open windows, and starting new creative projects (e.g., launching image editing software, creating idea notes).

Recent updates have particularly enhanced a feature called 'Copilot Vision.' This allows users to share parts of their PC screen or specific application windows with Copilot to receive context-aware guidance and operational support.

For example, features are provided such as 'Highlights,' where Copilot visually highlights and points out the relevant area on the screen when a user asks about how to perform a specific operation, and the ability to share up to two application windows simultaneously to analyze or ask questions across that information.

The OS-level integration of Copilot clearly demonstrates Microsoft's intention to place AI at the center of the daily computing experience. This has the potential to make AI assistance more accessible and intuitive for users.

On the other hand, since AI will always be running in the background, new challenges such as privacy considerations and user interface optimization will likely arise. It will be interesting to see how the fusion of OS and AI transforms the future PC user experience.

Copilot in Microsoft Edge (Browser Integrated Version)

Copilot in Microsoft Edge is an AI assistant feature built into Microsoft's web browser, Edge. It is typically easily accessible from the sidebar on the right side of the browser window and provides various features to enhance the web browsing experience.

The target audience is everyone who uses the Microsoft Edge browser. Key features include the ability to summarize the content of the web page currently being viewed and extract specific information. It is also possible to create new text based on the content of a web page or translate pages in foreign languages.

Copilot in Edge goes beyond simple text-based assistance. By utilizing the 'Copilot Vision' feature, you can share the screen you are viewing with Copilot and receive real-time responses and suggestions based on visual information.

Voice interaction is also supported, making it possible to ask questions or give instructions without using a keyboard. Furthermore, it is equipped with a feature that allows you to upload files such as PDFs or documents to Copilot to ask questions about their content or have them summarized.

Another distinctive feature is that users can select Copilot's response style from multiple modes, such as 'Balanced,' 'Creative,' and 'Deep Research,' depending on their needs and objectives.

From a data protection perspective, when using Copilot in Edge by signing in with a work or school Microsoft account (Microsoft Entra ID), enterprise data protection is applied, and the information entered and browsing history are handled according to the organization's security policy. Furthermore, companies can apply Data Loss Prevention (DLP) policies to Copilot in Edge to control the handling of sensitive information.

GitHub Copilot (for developers)

GitHub Copilot is an AI pair programmer designed to improve the productivity of software developers. While a developer is writing code, it suggests context-aware code autocompletion (ranging from entire lines to entire functions) in real time. This feature uses an AI system called OpenAI Codex developed by OpenAI, which understands context from existing code and comments to generate appropriate code.

The primary target audience is everyone who performs coding tasks, from professional software developers to students. It can be used with both personal and organizational accounts.

Key features include, in addition to the aforementioned code autocompletion, the ability to suggest code snippets based on requirements written in natural language (for example, in the form of comments in English). It supports numerous programming languages and major frameworks such as Python, JavaScript, TypeScript, Ruby, Go, C#, and C++, and also provides support for generating database queries.

It is worth noting that GitHub Copilot is provided free of charge to verified students, teachers, and maintainers of popular open-source projects. A 30-day free trial period is also available for general users.

In terms of delivery, it is provided as an extension for major text editors and integrated development environments (IDEs) such as Visual Studio Code, the JetBrains IDE suite, and Neovim, allowing developers to seamlessly receive support from GitHub Copilot within the environment they are accustomed to using.

The emergence and widespread adoption of GitHub Copilot have the potential to significantly change the way software is developed. Developers may be freed from routine code writing and repetitive tasks, allowing them to focus on more creative and complex problem-solving.

The free provision to students and open-source developers is a strategy aimed at expanding the user base and penetrating the community from an early stage, suggesting the possibility that it will become the standard for AI-assisted development in the future. On the other hand, new issues such as quality assurance of AI-generated code, copyright attribution, and security vulnerabilities have been raised, and these points require ongoing discussion and countermeasures.

Copilot for Specific Business Tasks

The Microsoft Copilot family, in addition to general-purpose AI assistant features, also develops solutions specialized for specific business areas and industry needs. These domain-specific Copilots aim to understand the workflows, data, and terminology unique to each domain to provide more accurate and practical assistance.

1. Copilot for Sales

Copilot for Sales is an AI assistant designed to help sales representatives improve their productivity and maximize results. It seamlessly integrates major customer relationship management (CRM) systems such as Microsoft Dynamics 365 Sales and Salesforce Sales Cloud with Microsoft 365 applications like Outlook and Teams, providing intelligent support in every aspect of sales activities.

Key features include meeting preparation (summarizing relevant customer information and past interactions), support for personalizing communication in emails and chats with customers, automation of activity logging and information updates in CRM systems, drafting proposals and emails, summarizing meeting content and extracting action items, and quick searching of internal and external information related to sales. This allows sales representatives to reduce time spent on administrative tasks and devote more time to building relationships with customers and strategic activities.

The price is an additional 20 USD per month for users who already have a Microsoft 365 Copilot license, and 50 USD per month for users who do not. Users of Dynamics 365 Sales Premium can access all features of Copilot for Sales by paying 30 USD per month for Microsoft 365 Copilot.

2. Copilot for Service

Copilot for Service is an AI assistant designed to improve operational efficiency and customer satisfaction for contact center operators and customer service representatives. It integrates with existing CRM systems and contact center solutions such as Dynamics 365 Customer Service, Salesforce Service Cloud, and ServiceNow to help accelerate and improve the quality of inquiry responses.

Key features include analyzing customer inquiries, quickly searching for relevant knowledge base articles and past similar cases, automatically generating draft responses for customers, summarizing interaction records, and providing real-time information to agents. This enables agents to resolve customer issues more quickly and accurately, and is expected to reduce training time and standardize response quality.

3. Copilot for Finance (for Finance Professionals)

Copilot for Finance is an AI assistant specialized for finance and accounting department personnel. It is expected to support tasks such as financial data analysis, automation of report creation, analysis of budget-to-actual variances, and identification of potential financial risks.

For example, it is expected to be used to perform complex analysis based on natural language instructions using financial data in Excel, or to generate drafts of monthly financial closing reports. We need to keep a close eye on future announcements for detailed features and pricing at the time of the official release.

4. Copilot for Security (Security Operations Support)

Copilot for Security is an AI assistant developed to support security operations personnel such as security analysts and IT administrators. It integrates and analyzes vast amounts of information such as alerts, logs, and threat intelligence from various security tools used by an organization (e.g., Microsoft Sentinel, Defender series), and streamlines security operations tasks such as threat investigation, incident response, and vulnerability management through natural language interaction.

Specific features include support for security incident triage, identification of attack patterns, analysis of impact scope, presentation of recommended countermeasures, automatic creation of security reports, and collection and analysis of the latest vulnerability information and threat trends. It also includes a pinboard feature to promote collaboration, facilitating smooth incident response within teams.

Unlike other Copilot products, the pricing structure adopts a consumption-based model based on the usage of computing resources called Security Compute Units (SCU). Provisioned SCUs are charged at a unit price of 4 USD per hour per SCU, and SCUs exceeding the provisioned amount due to increased demand are charged at 6 USD per hour.

The emergence of these specialized Copilots demonstrates the potential for AI not only to have general task-processing capabilities but also to understand more advanced and specialized knowledge and business processes, thereby accelerating digital transformation across various industries.

Through the automation of routine tasks and advanced data analysis, experts in each field are expected to be able to focus on more strategic and high-value-added work. However, to effectively utilize these AIs, a skill set for mastering AI in addition to specialized knowledge in each field is required.

Furthermore, establishing a process for humans to verify the validity of AI-generated analysis results and proposals, as well as addressing industry-specific regulatory and compliance requirements, will also be important challenges.

Microsoft Copilot Studio (Customization and Development Platform)

Microsoft Copilot Studio is an integrated development environment (IDE) that allows companies to extend the capabilities of Microsoft 365 Copilot to meet their specific needs, or to develop their own custom Copilots (AI chatbots or AI agents) using low-code or pro-code.

This enables the construction of AI solutions tailored to company-specific business processes and data sources that cannot be handled by the standard features of Copilot alone.

The primary target users are corporate information systems departments, developers, or power users in business departments who want to further deepen their use of Copilot.

The main features of Copilot Studio are as follows:

Building Custom Copilots
You can create dedicated AI assistants or chatbots from scratch that are connected to an organization's unique data sources (internal databases, business systems, knowledge bases, etc.) and business processes.

Extending Microsoft 365 Copilot
You can extend existing Microsoft 365 Copilot features using elements such as connectors, plugins, custom prompts, and topics. This allows you to add capabilities such as retrieving information by linking with specific internal systems or automating specific business workflows.

Low-code/Pro-code Development
It supports both intuitive low-code development using a graphical interface and more advanced pro-code development using Power Fx, JavaScript, etc., allowing for flexible customization according to the developer's skill level.

Creating and Managing AI Agents
You can design, deploy, and manage AI agents capable of executing complex tasks consisting of multiple steps.

As for the pricing structure, in addition to a plan that allows for up to 25,000 messages per month for 200 USD, a pay-as-you-go model (which requires a separate Azure subscription) is also available. Furthermore, if you hold a Microsoft 365 Copilot license, you may be able to use some features of Copilot Studio without an additional license.

The emergence of Copilot Studio promotes the "democratization of AI," allowing even those who are not AI development experts to relatively easily build AI solutions tailored to business needs.

Companies are expected not to be satisfied with uniform AI features, but to strategically customize and utilize AI to enhance their own competitive advantage.

While Copilot Studio can be a powerful tool for this purpose, the responsibility for quality control of customized AI, security governance, and the establishment of operational maintenance systems rests with the adopting company.

Features and Utilization within Microsoft 365 Apps

The true value of Microsoft 365 Copilot lies in its deep integration into major applications that many business professionals use daily, such as Word, Excel, PowerPoint, Outlook, and Teams, providing specific AI assistance tailored to the characteristics of each application.

This allows users to seamlessly leverage the power of AI without leaving their familiar work environment, which is expected to dramatically improve productivity. Below, we detail the specific features and usage of Copilot in major Microsoft 365 applications.

Copilot in Word: Document Creation, Editing, and Summarization

Copilot in Word is a powerful tool that supports every process of document creation. Key features include drafting based on prompts (instructions), editing and transforming existing documents, summarizing long-form content, and answering questions about document content through chat.

As a specific usage method, there is the "Drafting" feature. For example, by giving a specific instruction such as "Create a first draft of a press release for next week's new product launch, including the product's main features and target customers," Copilot will generate a base text.

Furthermore, it is possible to have it reference existing related files (such as product specifications or past press releases) saved on SharePoint or OneDrive to create a draft that is more contextually relevant.

Next, with the "Edit and Rewrite" feature, you can ask Copilot to change the tone of generated or existing text (e.g., "Change this proposal to a more formal style"), simplify redundant expressions, or correct grammatical errors. You can also utilize the "Summarize" feature to efficiently extract important information from long documents by giving instructions such as "Summarize the main points of this document in three bullet points."

Copilot also provides support for formatting. For example, you can automate tasks such as automatically creating tables from text data, formatting bulleted lists, or unifying heading styles throughout a document. A multilingual translation feature is also included to support global communication.

Furthermore, through "chat-based interaction" with Copilot, you can get answers to questions about the document's content (e.g., "What are the main risks mentioned in this chapter?"), advice for better expression (e.g., "How can I make this paragraph more persuasive?"), and even design-related advice such as font selection.

These features are available not only in the desktop version of Word but also in the mobile version, which is useful for document review and quick summary creation while on the go.

However, when using Copilot in Word, it is necessary to be aware of some limitations. Users have reported that reference links to information sources generated by AI may not be provided, the history of requests made once is not retained and cannot be checked later, and it cannot always respond to complex layout changes or advanced formatting instructions.

Given these limitations, it is important not to take Copilot's suggestions at face value, but for a human to ultimately review the content and make manual adjustments or corrections as necessary.

Copilot in Excel: Data Analysis, Formula Generation, and Visualization

Copilot in Excel aims to dramatically improve the efficiency of data work using spreadsheets. Key features include support for data analysis, suggestions for formula columns to perform complex calculations, discovery of insights through data visualization using charts and pivot tables, and highlighting, filtering, and sorting data based on specific conditions.

As a concrete example of its use, the first is the extraction of "data analysis and insights." For example, for a sheet containing sales data, simply instructing it to "display the sales trends by product category for the past three months in a line chart" will cause Copilot to generate an appropriate chart, helping you visually grasp data trends and anomalies.

Furthermore, you can also provide instructions in natural language for more complex analyses, such as "Identify the top 10% of customers by purchase amount in this customer list and calculate their average purchase price."

The "formula generation" feature unlocks Excel's powerful calculation capabilities for a wider range of users. For calculations that previously required complex combinations of functions or logical construction, you can simply instruct it with something like "Suggest a formula to calculate the total value of column C for rows where the value in column A is 100 or more and column B is 'Tokyo'," and Copilot will generate the appropriate formula. This is expected to reduce the time spent creating formulas and decrease calculation errors.

Regarding data "manipulation," you can automate tedious manual tasks with instructions such as "Highlight rows where the status column in this table is 'Completed' in green" or "Find cells in the registration date column where no date has been entered." Copilot also assists with converting to tables and basic formatting.

As a unique feature, you can also paste (or reference) images or charts inserted into an Excel sheet and ask Copilot questions about their content. This allows you to deepen your understanding through Copilot even for charts with unknown data sources or diagrams with complex content.

As a point of caution when using Copilot in Excel, for the features to work optimally, the target Excel workbook must be saved on OneDrive or SharePoint. Additionally, it is recommended that the data be formatted in a somewhat structured way, such as in a table format, so that Copilot can correctly interpret and analyze the data.

Copilot in PowerPoint: Presentation creation, design, and summarization

Copilot in PowerPoint assists in every stage of creating presentation materials, streamlining the process of turning ideas into visually compelling formats. Key features include automatic creation of new presentations based on topics or existing files, suggestions for slide content, assistance with design adjustments, summarization of long presentations, and Q&A regarding the material content.

As a concrete usage method, first, in "New Creation," Copilot automatically generates a draft that serves as the backbone of the presentation (title slide, agenda, main section slides, etc.) based on the topic or keywords specified by the user.

For example, you can provide instructions such as "Create a presentation on the latest trends in AI technology. The target audience is managers in the technical department, and please include key trends and their impact on business."

Even more powerful is the "Create from file" feature. If you specify an existing Word document or PDF file (in the case of the Microsoft 365 Copilot version), Copilot will analyze its content and automatically convert it into presentation slides.

In recent updates, a feature has been added to reference up to five different files (Word, PDF, TXT files, etc.) and integrate that information to create a single presentation, significantly improving the efficiency of information gathering and material creation.

For created presentations or existing materials, you can use the "Summarize" feature to extract key points, or use the "Edit/Add Content" feature to add speaker notes to specific slides, organize the slide structure, or suggest and add images that match the theme. For example, specific editing instructions such as "Add a sentence to the conclusion of this slide that encourages action" are also possible.

In terms of design, there is a "Branding" feature that is particularly important for corporate use. You can have slides generated based on official brand templates or color schemes established by the company, which helps maintain a consistent, professional appearance. You can also instruct Copilot to "translate" the entire presentation.

As a note on usage, a separate Microsoft Designer license may be required to use some advanced features such as adding slides or creating presentations from files. Additionally, Copilot in PowerPoint supports voice commands, allowing you to provide instructions verbally via the microphone button.

Copilot in Outlook: Email management, draft creation, and coaching

Copilot in Outlook is an AI assistant that helps streamline daily high-volume email tasks and improve the quality of communication. Key features provided include quick summarization of received email threads, creation of new email drafts or generation of reply suggestions based on the purpose, adjustment of the tone of written emails, and AI-driven coaching to help create more effective emails.

As a concrete usage method, there is first the "Email Summary" feature. When you receive long email threads with many exchanges or individual long emails with a large amount of information, Copilot automatically extracts the key points, major decisions, and action items, and presents them in a concise summary.

In recent updates, you can now summarize the content of Word, PowerPoint, and PDF files attached to emails directly without leaving Outlook. This reduces the risk of missing important information and significantly shortens the time spent processing emails.

Next, with the "Draft Creation" feature, Copilot generates an email draft simply by the user providing a brief instruction (prompt). For example, specific instructions such as "Create an email to Mr./Ms. B at Company A requesting to reschedule next week's meeting. Please use polite language and suggest three candidate dates from our side" are possible.

　Once a draft is generated, you can instruct Copilot to "adjust the tone or edit" it as needed. For example, you can change it to a more formal tone, make it more casual, or shorten or lengthen the overall text.

　A unique feature is "Coaching." This allows Copilot to provide improvement suggestions from an AI perspective on email drafts created by the user. Specifically, it offers concrete advice and revisions regarding whether the tone of the message matches your intent, whether the content is clear, and how the recipient might interpret it (sentiment analysis). This helps prevent communication errors and fosters more effective and professional email writing skills.

　Furthermore, it includes a "Translation" feature that supports multilingual communication, assisting with translating received emails and drafting emails in foreign languages.

Copilot in Teams: Meeting summaries, action items, and chat assistance

　Copilot in Teams is a powerful AI assistant designed to improve the efficiency and quality of online meetings and team communication. It provides a wide range of features, such as organizing information during and after meetings, and supporting quick information sharing and decision-making in chats.

Its use in meetings is particularly noteworthy.

During meetings

Real-time summarization
　Even while a meeting is in progress, you can instruct Copilot to provide a real-time summary of the key discussion points so far (who said what, points of agreement or disagreement, etc.). This allows participants who joined late or whose focus was temporarily interrupted to quickly grasp the flow of the discussion.

Action item suggestions
　Copilot identifies specific tasks and owners from the discussion and suggests listing them as action items.

Q&A and issue organization
　You can ask Copilot questions about points raised during the meeting, have it identify unresolved issues, or organize different participants' perspectives on specific topics. For example, you can ask, "What were the main concerns of Person A regarding this proposal?"

Meeting minutes generation support
　Copilot assists in generating a draft of meeting minutes based on what was said during the meeting.

After meetings

Comprehensive understanding of meeting content
　Based on meeting chat history or recordings (transcribed transcripts), Copilot provides comprehensive information and insights about the entire meeting. For example, you can give instructions like, "Summarize the main ideas discussed in this meeting and each participant's reaction to them."

Follow-up support
　You can also have Copilot create a draft of a follow-up email that includes the meeting summary, decisions made, and action items.

Use in chat

Catching up on chat content
　In chat channels or direct messages where many messages are exchanged, Copilot quickly summarizes unread parts or key points of conversations, action items, and decisions regarding specific topics. For example, you can efficiently obtain information by asking, "What did Person A say in the chat yesterday about the progress of Project B?"

Message drafting support
　You can ask Copilot to rewrite or adjust the tone of messages you send in chat.

　Please note that to use many of the features of Copilot in Teams, especially those related to meetings, meeting transcription must be enabled. However, depending on the meeting organizer's settings, some Copilot features may be available during the meeting even without transcription.

Copilot in OneNote: Note organization, summarization, and idea generation

　Copilot in OneNote is an AI feature that supports organizing, understanding, and creating information within OneNote, a digital note-taking application. It analyzes the content of handwritten or typed notes and supports a wide range of tasks, such as creating summaries, extracting task lists, analyzing and generating ideas, and even drafting various plans. It is also possible to respond to questions about the content of your notes in real-time through a chat interface.

As a concrete example of how to use it, support for "understanding notes" is a primary use case. You can have Copilot generate summaries for entire OneNote sections, specific pages, or selected text ranges. For example, for notes recorded in a rambling manner during a long meeting, you can give instructions such as "Summarize the key decisions and next actions from these meeting notes in a bulleted list" to organize information and quickly grasp the important points.

Next, the "task list creation" feature automatically extracts tasks that need to be executed from your notes and organizes them into a to-do list. This prevents you from missing action items that might otherwise get buried during brainstorming or discussions.

For example, you could give instructions such as "From the notes of today's brainstorming session, list the tasks that require follow-up, and include proposed assignees and deadlines."

The "idea generation" and "drafting plans" features are particularly useful for creative work or the early stages of a project. When you want ideas for a new project or are planning an event, you can use prompts like "Suggest 5 ideas for a new marketing campaign.

The target audience is young people in their 20s" or "Create a draft plan for the internal training session to be held next month. Please include the purpose, target audience, main content, and schedule" to get concrete proposals or outlines from Copilot.

It is worth noting that Copilot in OneNote also supports handwritten notes. By selecting handwritten notes and giving instructions to Copilot, you can perform summaries, task extraction, and content analysis just as you would with typed notes.

In a recent update, a feature called "Copilot Notebooks" was introduced within OneNote. This allows you to aggregate relevant content from various sources (OneNote notes, Word documents, websites, meeting recordings, etc.) into a single notebook, and Copilot understands and analyzes the entire aggregated information to suggest deeper insights or highly relevant actions.

Copilot Notebooks re-scans in real-time when information is updated, providing support based on the latest situation at all times.

Copilot in Loop: Collaborative Content Creation and Summarization

Microsoft Loop is an application that allows teams to organize information by combining components (text, lists, tables, etc.) on a flexible canvas and collaborate in real-time. Copilot in Loop aims to bring the power of AI into this collaborative environment, making processes such as content creation, editing, and summarization more efficient and creative.

As for specific features and usage, there is "content creation support" within Loop pages. Users can give instructions to Copilot in natural language to generate new text blocks, lists, or drafts of ideas.

It is also possible to select existing content and perform "rewriting of existing content" or improvements with instructions such as "Rewrite this paragraph to be more concise" or "Add potential risks to this list."

When a Loop page becomes long or the information is diverse, the "page summary" feature is useful. By instructing Copilot or selecting it from the "/ (slash)" command menu on the page, you can automatically generate and insert a summary that concisely summarizes the main points of the entire page.

In Loop, where collaborative editing by teams is active, it is important to grasp the changes made by other members. The "Recap" feature of Copilot in Loop allows Copilot to automatically capture updates made by users to a page and display them as a summary in the Recap pane. This allows team members to quickly grasp the latest situation and prevent duplication of work or oversights.

Furthermore, when linking Teams meetings and Loop, a feature is provided where Copilot automatically captures key discussion points and proposed tasks as "collaborative notes" during the meeting and records them on the Loop page. For example, specific instructions such as "Record the feedback from the customer mentioned by Person A as a quote" are also possible.

To use Copilot in Loop, you generally need a Microsoft 365 Copilot license. Users without a license can view or limitedly edit content created by other license holders using Copilot, but they cannot give direct instructions to Copilot or engage in dialogue with the AI.

AI integration into collaborative tools like Copilot in Loop has the potential to contribute significantly to improving team productivity. However, to maximize its effectiveness, all team members need to understand the capabilities and limitations of AI, and learn how to provide effective prompts and appropriately evaluate and utilize content generated by AI.

The fact that Microsoft 365 Copilot is beginning to link more seamlessly between applications is a trend worth noting. For example, features that allow you to create presentations by referencing multiple files (Word documents, PDFs, text files, etc.) in PowerPoint, or summarize email attachments (Word, PowerPoint, PDF) directly within Outlook, significantly improve efficiency when users work across applications. Features like Copilot in Loop, which assume that multiple users will perform real-time collaborative editing with the support of AI, also accelerate this trend.

These evolutions suggest that Copilot is evolving from limited work support within a single application to intelligently supporting more complex workflows across multiple applications and services.

In the future, a series of tasks such as Copilot drafting a report in Word based on meeting content in Teams, summarizing the main points in PowerPoint slides, and drafting a report email to relevant stakeholders in Outlook may be able to be performed more smoothly and automatically.

However, to realize such advanced cross-application integration, it is essential to further improve the context understanding capabilities of the Copilot embedded in each application and to further strengthen the data integration mechanisms between applications.

Furthermore, ensuring transparency so that users can clearly understand and appropriately manage which data is being used by Copilot, in which application, and how, will also be extremely important from the perspective of improving reliability.

Introduction and Utilization of Microsoft Copilot

To maximize the use of Microsoft Copilot's wide range of features, it is essential to select the appropriate edition, meet the prerequisites, and master effective usage methods. This section explains the licensing and pricing structure, the system requirements and setup procedures necessary for implementation, and the basics and best practices of prompt engineering to master Copilot.

Licensing, Pricing, and Subscription Models

Microsoft Copilot has a wide variety of editions, licenses, and pricing, ranging from free versions to paid versions for individuals, businesses, specific tasks, and development platforms.

Free Version of Microsoft Copilot
Basic AI assistant features (chat, search, content generation, etc.) are available for free. By signing in with a work or school Microsoft account, features that protect input data for commercial use are enabled.

Microsoft Copilot Pro
A paid version aimed at improving individual productivity, priced at 3,200 yen or 20 USD per user/month. It provides priority access to the latest AI models and integration features with Microsoft 365 Personal/Family subscription apps (Word, Excel, PowerPoint, etc.).

Microsoft 365 Copilot (Business/Enterprise)
The most comprehensive plan for businesses, with a base price of 30 USD per user/month (annual contract). This requires a separate, qualifying Microsoft 365 license (e.g., Microsoft 365 Business Standard/Premium, Microsoft 365 E3/E5, etc.) as a prerequisite.

GitHub Copilot
An AI pair programmer for developers, with multiple plans. These include Copilot Free (limited features), Copilot Pro (10 USD/month or 100 USD/year), Copilot Pro+ (39 USD/month or 390 USD/year), Copilot Business (19 USD/user/month), and Copilot Enterprise (39 USD/user/month). Verified students, teachers, and maintainers of popular open-source projects may be able to use it for free.

Copilot for Specific Tasks (Copilot for Sales/Service, etc.)
For example, Copilot for Sales and Copilot for Service each have a base price of 50 USD per user/month. If you already hold a Microsoft 365 Copilot license, you can add these specialized features for an additional fee (e.g., 20 USD per user/month).

Copilot for Security
A Copilot specialized for security operations, which adopts a pay-as-you-go model based on the consumption of computing resources called Security Compute Units (SCU). Provisioned SCUs cost 4 USD per hour per SCU, and excess SCUs cost 6 USD per hour.

Microsoft Copilot Studio
A platform for building custom Copilots and AI agents, with plans starting at 200 USD per month (up to 25,000 messages) and pay-as-you-go options using Azure subscriptions.

As shown, the pricing structure for Copilot is very diverse, and the total cost varies significantly depending on the edition selected and the status of existing Microsoft licenses. When considering implementation, it is important to comprehensively consider your organization's or individual's needs, the features you want to use, and your budget to select the optimal plan. Especially for business plans, it is necessary to also take into account the cost of the underlying Microsoft 365 license.

Prerequisites: System Requirements and Setup

To implement Microsoft Copilot, especially Microsoft 365 Copilot for business, and to maximize its functionality, several prerequisites must be met and appropriate setup must be performed. Neglecting these preparations may result in limited Copilot functionality or failure to achieve the expected performance.

Main Prerequisites for Microsoft 365 Copilot

Overview of Setup Checklist 10

To ensure a smooth implementation, it is recommended to verify the following steps.

These prerequisites are wide-ranging, and since areas such as network settings, privacy settings, and license management require specialized knowledge, it may be difficult for individual users to handle them.

Therefore, when introducing Microsoft 365 Copilot in a corporate organization, it is essential for the information systems department to take the lead and assume roles such as planned preparation, deployment, configuration, and post-implementation troubleshooting.

Clearing these prerequisites at the initial stage is the key to maximizing the benefits of Copilot implementation. The IT department bears the important responsibility of establishing an environment where employees can utilize Copilot's capabilities smoothly and safely while ensuring security and compliance.

Evaluation of Microsoft Copilot: Benefits, Limitations, and Ethical Considerations

Microsoft Copilot holds great potential for leveraging AI to improve individual work efficiency and corporate productivity; however, when introducing and utilizing it, it is necessary to comprehensively evaluate not only the benefits but also the potential limitations, risks, and ethical aspects.

Key Benefits: Improved Productivity, Efficiency, and Creativity

The main benefits expected from introducing Microsoft Copilot are a dramatic increase in productivity, improvement in operational efficiency, and the further unleashing of human creativity.

Increased Productivity
Copilot can automate or semi-automate routine daily tasks such as drafting emails, creating report drafts, assisting with meeting minutes, and data entry. This frees employees from cumbersome tasks, allowing them to increase the time spent on more strategic, high-value, and creative work.

Time Savings
You can significantly reduce the time spent searching for information on the internet or internal databases, reading long documents (by utilizing Copilot's summarization features), and creating presentation materials or reports. This enables shorter project lead times and the ability to handle more tasks.

Decision-Making Support
By linking with applications like Excel, it assists in analyzing large amounts of data, visualizing trends, and extracting insights. This makes it possible to make faster and more accurate data-driven decisions.

Improved Communication
Through email drafting assistance in Outlook (such as tone adjustment and suggestions for appropriate expressions), summarizing key points of meetings in Teams, and multilingual translation features, you can improve the quality and efficiency of communication both inside and outside the company.

Stimulating and Expanding Creativity
By assisting with brainstorming new ideas, suggesting content from diverse perspectives, or using image generation features, it helps stimulate human creativity and generate new concepts.

Seamless Integration with the Microsoft Ecosystem
One of Copilot's major strengths is its deep integration into the Microsoft 365 application suite that many users utilize daily, such as Word, Excel, PowerPoint, Outlook, and Teams. This allows users to naturally receive AI assistance in an environment they are already familiar with, keeping the learning cost low when introducing new tools and facilitating smooth adoption and utilization.

Taken together, these benefits suggest that Microsoft Copilot has the potential to contribute not only to increasing individual work efficiency but also to improving overall organizational productivity, promoting innovation, and strengthening competitiveness.

Limitations, Risks, and Security Concerns

Behind the many benefits provided by Microsoft Copilot, there are limitations, potential risks, and security concerns that require special attention. Understanding and appropriately addressing these is the key to using Copilot safely and effectively.

Accuracy Issues
AI based on LLMs does not always generate 100% accurate information. Information presented by Copilot may contain factual errors, incomplete data, or content that is out of context. In particular, the phenomenon known as "hallucination," where AI generates plausible-sounding falsehoods, can affect important decision-making. Therefore, it is essential not to take Copilot's output at face value but to perform human fact-checking and critical examination. The importance of this AI literacy will only increase as Copilot becomes more widespread.

Limitations in Context Understanding
While Copilot understands the context of a conversation to some extent, it can be difficult for it to fully grasp highly complex nuances, tacit knowledge in specialized fields, or background information not explicitly stated in the prompt. Additionally, since conversation history is not permanently retained in some Copilot sessions, users may need to reorganize information themselves for long conversations or complex tasks.

Intellectual Property Rights
There is a risk that in the process of generating content based on training data, Copilot may unintentionally produce expressions or ideas similar to existing copyrighted works, thereby infringing on intellectual property rights. While Microsoft has indicated a policy of protecting users from legal risks under certain conditions through the "Copilot Copyright Commitment," users still need to exercise careful verification and caution when using AI-generated content commercially.

Data Privacy and Security
Microsoft 365 Copilot, in particular, accesses data that may contain sensitive information, such as emails, documents, and chats within an organization, through Microsoft Graph. Therefore, the risk of this information being accessed unauthorizedly or leaked externally is not zero.

Microsoft has implemented measures such as data encryption, strict access control, inheritance of existing Microsoft 365 security policies, and a commitment not to use user data for general training of AI models.

However, risks such as inadequate access rights settings within an organization or misuse by malicious insiders still remain. To minimize these risks, it is essential for the organization to establish and operate an appropriate data governance framework.

Impact on Brand Messaging
When Copilot assists in creating official corporate documents or customer communications, the text or expressions generated by the AI may not always perfectly align with the company's brand image or communication strategy. To avoid the possibility of undermining brand consistency, AI-generated content requires careful review and adjustment by humans.

Possibility of Over-reliance
Due to the convenience of Copilot, there is a risk that users may uncritically accept AI suggestions or generated output, failing to fully exercise their own critical thinking and professional judgment. It is important to always keep in mind that AI is merely a support tool, and the final decision-making and responsibility should rest with humans.

Cost
For individual users and small to medium-sized businesses in particular, the adoption cost of paid versions like Copilot Pro or Microsoft 365 Copilot can be a burden. It is necessary to carefully consider the cost-effectiveness and select a plan that meets your needs.

Integration Constraints
While Copilot is optimized for collaboration within the Microsoft ecosystem, there may be limitations when integrating with non-Microsoft business systems or special data formats used by an organization.

Learning Curve
Although Copilot is designed to be intuitive, fully mastering its diverse features and advanced usage methods (such as creating effective prompts) may require a certain amount of learning and trial and error.

Fully understanding these limitations and risks while implementing and utilizing Copilot is the key to maximizing its benefits.

Responsible AI: Microsoft's Approach and User Responsibility

With the spread of powerful AI technologies like Microsoft Copilot, consideration for their ethical aspects becomes extremely important. Microsoft has set forth the principles of "Responsible AI" and has shown a stance of addressing ethical issues in technology development and provision, but at the same time, corresponding responsibility is required from the users who utilize AI.

Microsoft's Ethical Principles and Initiatives

Microsoft has established the following six key ethical principles to ensure that AI systems benefit society and are used in a trusted manner: Fairness, Reliability and Safety, Privacy and Security, Inclusiveness, Transparency, and Accountability.

Bias Mitigation
AI models can reflect biases contained in training data. Microsoft is making efforts to detect and mitigate unintended biases through ensuring fairness in algorithm design and continuous monitoring of training data and AI models. However, it is extremely difficult to completely eliminate all biases that exist in society, and efforts to address this challenge are ongoing.

Transparency and Accountability
Microsoft is striving to increase transparency as much as possible regarding how Copilot operates and what data it uses to generate suggestions and content. Furthermore, it values user feedback and aims to fulfill its accountability by investigating the causes and leading to improvements if there are issues with the AI's behavior.

Data Privacy and Security
As mentioned above, Microsoft is committed to ensuring privacy and security through strict management of user data, the application of encryption technology, thorough access control, and compliance with international data protection regulations such as the GDPR (EU General Data Protection Regulation). Specifically regarding Microsoft 365 Copilot, Microsoft has clearly stated that customer organization data is not used to train external general-purpose AI models.

Consideration for Employee Well-being
Care must be taken to ensure that the introduction of AI technology, especially AI that may have aspects of a monitoring tool, does not cause excessive stress or pressure on employees. Microsoft has expressed the view that Copilot is a tool to augment and support employee capabilities, and it should not hinder the way they work.

User Responsibility

The ethical use of AI technology cannot be achieved solely through the efforts of the developers. Individual users and organizations that use Copilot also bear the following responsibilities.

Critical Evaluation and Final Judgment
Rather than uncritically accepting information or suggestions generated by AI, it is necessary to always examine the content and make final judgments based on one's own expertise and experience.

Compliance with Ethical Use Guidelines
As an individual and as a member of an organization, one should strive to understand and comply with guidelines regarding the ethical use of AI (whether established by the organization or as social norms).

Appropriate Handling of Confidential Information
When inputting highly confidential information into Copilot, it is necessary to follow the organization's data security policy and take care to minimize the risk of information leakage.

Reporting Inappropriate Behavior
If Copilot provides biased information, generates inappropriate content, or exhibits other ethically problematic behavior, reporting it to Microsoft through feedback will lead to the improvement and healthy development of AI.

Best Practices for Responsible AI Deployment

When companies introduce and deploy Copilot, it is recommended to practice the following best practices.

Formulate clear internal policies that define the scope of Copilot usage, prohibited items, review processes, and procedures for reporting concerns.

Regularly conduct ethics-based audits that evaluate the fairness, accountability, and transparency of AI systems.

Build and operate a strong data governance system that includes access control to confidential data, data retention/deletion policies, and real-time auditing.

Provide education and training to employees not only on how to operate Copilot, but also on ethical considerations, situations where human judgment is required, and escalation procedures when problems occur.

Appoint AI ethics officers and clarify the supervisory roles of relevant departments such as IT and Human Resources.

In conclusion, the ethical realization of advanced AI like Microsoft Copilot can only be achieved through the cooperation of both sides: the technical efforts of Microsoft as the developer, and the understanding, literacy, and active involvement of the humans who use it in a way that complements the AI.

Moving forward in the AI society, it will become increasingly important to have a perspective of "how to collaborate to produce better results" rather than simply viewing AI as a target to which "work is delegated."

Key New Feature Highlights of Microsoft 365 Copilot (as of May 2025)

Copilot Notebooks (OneNote)
It is now possible to use "Copilot Notebooks" in OneNote. Users can aggregate various content such as notes, documents, the web, and meeting minutes into a single notebook, and based on this, Copilot provides optimal summaries and insights while always referring to the latest information. It also features the ability to periodically summarize the contents of the notebook via audio (scheduled for rollout in June 2025).

Copilot Search Result Summarization (Edge Browser)
When performing an in-page search (Ctrl+F) in the Edge browser, a feature will be added where Copilot displays a summary of content related to the search term. When you want to quickly grasp information within a web page, you can check the main points summarized in the sidebar (scheduled for rollout in June 2025).

Inserting Citations & Voice Interaction in Word
　A feature has been added where Copilot automatically inserts source information as citations within Word documents. This makes it easier to create documents that clearly state their sources (scheduled for rollout in June 2025). Additionally, you can now engage in voice interaction with Copilot in Word, allowing you to ask questions about the entire document or specific sections to deepen your understanding, or provide instructions for content corrections via voice (available in the Web and Windows versions of Word starting in May 2025).

Long-form Summarization and Coaching Features (Word)
　A "Document Snapshot" feature has been added to dynamically summarize long documents, allowing you to select the level of detail for the summary (short, medium, or detailed). Furthermore, a writing guidance feature called Coaching with Copilot has been included, providing improvement advice that goes beyond simple grammar checks, such as suggestions for writing style, structure, and additional information. Users can apply these suggestions to their text with a single click (available starting May 2025, initially in English).

　Example of the coaching feature for Word documents using Copilot. It analyzes the text and provides professional improvement suggestions, such as "change to a softer tone for criticism."

Enhanced Presentation Creation Support in PowerPoint
　Copilot can now reference up to five files (existing materials, PDFs, or TXT documents) in PowerPoint to automatically generate presentation materials from a wealth of information sources (available May 2025).

　For example, you can specify a report PDF and text requirements together, and have them reflected in slides all at once. Also, when creating new materials from existing slides using the narrative construction feature, it has been improved to create a separate file rather than editing the original. A feature to start generating a new presentation directly with Copilot from the PowerPoint file menu (Backstage view) has also been added.

Expansion of Copilot Chat
　Microsoft 365 Copilot Chat (a feature for asking Copilot work-related questions in a chat format) has also been enhanced. You can now schedule recurring executions for frequent routine questions, save and customize prompts according to your own work processes, and search and reuse past chat history.

　For example, automation such as "summarize sales reports every morning at 9 AM" is now possible. This feature became available in May 2025. Furthermore, features for voice input in Copilot Chat and having Copilot read its answers aloud have been added (available in June and May, respectively), improving hands-free usage and accessibility.

Attachment Summarization in Outlook
　A feature has been added where Copilot automatically summarizes the content of Word, PowerPoint, or PDF documents attached to Outlook emails and displays it directly within the email reading pane. You can grasp the gist without having to open the attachments one by one, and you can even ask follow-up questions about the summary content (available starting May 2025).

Examples of Business Utilization　
With the new features mentioned above, Microsoft 365 Copilot is increasingly being used for business automation and document creation support. For example, Copilot can also be used for meeting minutes and extracting to-do items.

　Copilot for Microsoft Teams (Teams Copilot) summarizes meeting content in real-time and presents the responsible parties and action items that require follow-up, helping to prevent important matters from being overlooked even in busy meetings.

　When creating reports in Word, Copilot generates a draft, and users can significantly reduce the time spent on document creation by adding to and editing it. Additionally, since sources (citations) are automatically inserted into Copilot's suggestions, it also streamlines evidence verification when writing reports.

　In Excel as well, Python integration has been progressing since the end of 2024, and features have been enhanced for Copilot to execute advanced data analysis and forecasting using Python based on natural language instructions. This allows even those without specialized knowledge to perform statistical analysis or create graphs by asking questions like "create a sales forecast for next year."

　Furthermore, by utilizing the OneNote Copilot notebook, it is possible to consolidate various project materials across the board to organize key points and quickly review them via voice. In this way, Microsoft 365 Copilot is evolving to comprehensively support business tasks, from daily information organization to material creation and email correspondence.

Evaluation by Adopting Companies
　Since the preview in 2023, expectations for Microsoft 365 Copilot have been rising, especially among large enterprises, and organizational preparation for adoption is underway, with 90% of Fortune 500 companies participating in agent development using Copilot Studio (described later).

　For example, some companies are conducting trials using Copilot to quickly extract necessary information from internal documents or to generate drafts for proposals and emails, with feedback reported such as "the time for draft creation has become less than 1/3 of what it used to be" and "follow-ups after meetings are no longer missed" (*specific company names may be withheld).

　In particular, companies participating in the Frontier Program (an early access program for Microsoft 365 Copilot users) have evaluated that business efficiency is improving dramatically due to the continuous evolution of Copilot.

Key New Feature Highlights of GitHub Copilot (up to May 2025)

Vision for Copilot (UI Code Generation from Image Input)
　A "Vision" feature has been added to Copilot Chat that allows you to upload images (screenshots or hand-drawn UI mocks) and automatically generate corresponding code or designs based on their content.

　For example, if a marketing representative writes correction instructions in red pen on a screenshot of a web page and uploads it, Copilot will generate the HTML/CSS code reflecting those changes and even the alt text for images. This feature was initially provided as an extension for VS Code, but it is now natively integrated into GitHub Copilot Chat and can be used without an API key.

Next Edit Suggestions
　Copilot can now predict and suggest "the next changes to make" based on the code edits the user has recently performed. While conventional Copilot completion was limited to the context around the cursor position, the new algorithm considers the edit history of multiple previous lines and displays candidates looking one step ahead, such as "If you are updating this function next, shouldn't you also add arguments?" If you like the suggestion, you can accept it with a single Tab key press, or ignore it if it is unnecessary. It has evolved into smarter code completion that understands context better.

Multi-file editing (Copilot Edits) and agent mode
The "Copilot Edits" feature, which allows for bulk code changes across multiple files and was released in preview in 2024, has now been officially released. If you give an instruction in natural language such as "Change the user ID type from int to GUID," it will detect all relevant files and edit the corresponding sections at once.

Furthermore, an agent mode has been added, allowing Copilot to automatically identify the group of files related to the intent of a request and apply changes without the user having to specify each file to be edited one by one. GitHub CEO Thomas Dohmke also commented that this is "a step closer to Copilot understanding the user's intent and autonomously planning and executing the necessary work."

Code review support and Issue creation
The feature where Copilot assists with Pull Request code reviews on GitHub is being expanded. The Copilot Code Review feature, which reads code diffs to summarize key points and provides suggestions pointing out improvements or potential bugs, has been improved, refined for better coverage, and polished with a UI that is easier for developers to control. Additionally, a feature to have Copilot generate Issues on the GitHub web interface was released in preview in May 2025, allowing for the automatic registration of bugs and improvements within a repository through natural language dialogue.

External AI model support and usage limits
While Copilot is based on OpenAI's GPT-4 series models, it is now possible to use advanced external LLMs such as Anthropic's Claude 3.7/4. On the other hand, a "Premium Request" system has been introduced that sets an upper limit on the number of monthly requests when using advanced models. While the standard base model (the original fine-tuned GPT-4o) is unlimited, using higher-precision models like Anthropic's requires, for example, up to 300 requests per month on the Copilot Pro plan, with pay-as-you-go pricing ($0.04 per request) or a higher-tier plan required beyond that. This is a measure taken because high-precision models consume significant computational resources, but it is also against the backdrop of extremely high demand for GitHub Copilot itself, with Copilot driving over 40% of GitHub's total revenue growth in 2024.

Examples of use in the engineering domain
GitHub Copilot is becoming established as a "pair programmer" in software development environments, and the following specific uses can be seen.

Automatic code generation and completion
Developers can generate the skeleton of code instantly just by describing the outline of a function or algorithm in comments. In daily coding, the time-saving effect of automatic generation of boilerplate code is significant, and one company has even reported that "Copilot generates about 40% of developers' code" (official GitHub announcement). With the new "Next Edit Suggestion" feature, preventing omissions during refactoring (proposing additional changes to related areas) has also become more efficient.

Bug fixing and debugging support
When you ask questions using Copilot Chat such as "What is the cause of this error?" or "Review this function for bugs," it analyzes the relevant code and suggests the cause of the bug or improvement plans. Furthermore, with the enhancement of the code review feature, Copilot summarizes changes and points out potential issues when creating a Pull Request, contributing to improved code quality.

Automatic generation of documentation and tests
To allow developers to focus on implementation, Copilot also generates function docstrings (documentation comments) and unit test templates. If you briefly state the purpose of a function, a docstring is proposed, and in some cases, Copilot will ask, "Would you like to add a test case corresponding to this change?" in response to code modifications. This is expected to improve test coverage and reduce the effort required for documentation maintenance.

DevOps integration and automation
Collaboration between the GitHub CLI and Copilot, as well as utilization within CI/CD platforms incorporating Copilot, is also progressing. For example, it is possible to instruct Copilot in natural language to generate YAML configuration files for CI or create operational scripts.

In 2025, a preview of "Copilot Agents" was introduced, and initiatives have begun where, when developers provide ambiguous requirements, the agent autonomously carries out multiple tasks (code editing, test execution, deployment configuration, etc.). Looking to the future, scenarios are being envisioned, such as GitHub's "Project Padawan," where AI performs the entire implementation process just by a developer submitting an Issue.

Evaluation by adopting companies
GitHub Copilot is already licensed for developers at many companies, and there are many voices saying that "development speed has improved dramatically." For example, one software company reported that after introducing Copilot, code review comments decreased and the time to release was shortened.

Acceptance of Copilot is also progressing in the OSS community, and the number of VS Code extension downloads and Copilot users continues to increase. In addition, with the free version of Copilot (a feature limited to public GitHub repositories) that became available at the end of 2024, it is also spreading to startups and individual developers.

While some users were initially confused by the fact that using advanced models costs money, opinions such as "the flexibility to choose cutting-edge models as needed is welcome" have emerged, as well as operational examples of balancing cost and performance through appropriate request management.

Main new feature highlights of Power Platform Copilot (up to May 2025)

Power Platform Copilot is a collective term for AI-assisted features provided on the Microsoft Power Platform, such as Power Apps, Power Automate, Power Virtual Agents, Power BI. It plays a role in helping citizen developers and business department users build apps and workflows without writing specialized code.

As of May 2025, Copilot is deeply integrated into each Power Platform service, realizing the automation of business processes and rapid solution development.

Natural language app development in Power Apps
In Power Apps, a feature that automatically generates data models and screens just by describing requirements on the canvas in natural language has been made generally available. When business requirements are entered on the integrated development canvas, React code is generated in the background, and a functional app page is built immediately.

For example, if you enter "I want to create an app to manage customer inquiries. I need screens for a customer list and response status," the relevant data tables (customer entities) and UI screens are proposed and created by Copilot. This has made it possible for even non-professional developers to prototype business apps in a short time.

Automatic flow creation in Power Automate
In Power Automate, if a user dictates steps such as "I want to create a flow to email a summary report to the sales team every Monday," Copilot will build the entire workflow.

Furthermore, in 2025, Copilot introduced an "autonomous agent" feature that utilizes a library of triggers and actions to monitor common business events such as email receipt or SharePoint updates and respond automatically.

For example, an AI agent can execute a series of processes without human intervention, such as "automatically replying when an inquiry arrives from a specific email address and recording the content in a database."

Chatbot Generation with Power Virtual Agents
Copilot is also integrated into Power Virtual Agents, a chatbot building service. Users can simply provide instructions such as "Create a bot for the internal IT help desk. Make it able to answer questions about password resets and VPN connection methods," and a chatbot will be created with automatically generated relevant QA pairs.

By having Copilot extract knowledge from corporate FAQs and existing documents and incorporate it into the bot, the time required for bot development can be significantly reduced. In 2025, a feature for this bot creation that coordinates with other agents to supplement knowledge is also being previewed.

Copilot Studio and Multi-Agent Coordination
At Microsoft Build 2025, new features for "Copilot Studio" were announced. Copilot Studio is a platform that allows companies to easily create and tune their own unique Copilots (agents).

With the new "Copilot Tuning" feature, you can fine-tune AI models to your company's data and business processes without programming, allowing you to create domain-specific Copilot agents.

Furthermore, multi-agent orchestration has become possible, allowing multiple agents to link together and work like a team. For example, a scenario can be realized where agents created for the HR, IT, and Marketing departments collaborate to share the tasks involved in the new employee onboarding process.

An agent store is also available on Copilot Studio, where you can add third-party agents such as Jira, Monday.com, and Miro, as well as your own custom agents.

Integration with Azure AI Foundry
Copilot Studio has also strengthened its integration with Azure AI Foundry (a professional AI model development environment). For example, you can import industry-specific AI models (over 1,900 in total, including those for healthcare and finance) to improve the response accuracy of your Copilot agents.

It also supports "Bring Your Own Model," which allows you to incorporate non-Microsoft models, enabling companies to freely select and integrate the models best suited to their requirements. These are currently available as a public preview.

Key New Feature Highlights of Dynamics 365 Copilot (as of May 2025)

Dynamics 365 Copilot is an AI assistant feature integrated into Microsoft's CRM/ERP product suite (including sales, marketing, customer service, field service, finance and accounting, and supply chain management). As of May 2025, Copilot has been enhanced across all Dynamics 365 applications to provide industry- and task-specific support.

Sales (Dynamics 365 Sales)
Copilot features that intelligently support sales activities, such as automated research, follow-ups, and suggestions for priority tasks, have been enhanced for sales representatives. For example, Copilot analyzes emails and call content with customers and presents "next actions to take" (such as sending additional proposals or reporting to a supervisor) without fail. The UX of the sales dashboard has also been revamped, allowing for pipeline management with Copilot's advice naturally integrated.

Customer Service (Dynamics 365 Customer Service / Contact Center)
In support operations, Copilot assists with case handling and knowledge management like an agent. The AI automatically determines the intent from the inquiry content and immediately presents appropriate response templates or relevant knowledge articles, or proposes solutions directly when necessary.

In the 2025 update, this automation of intent inference and knowledge retrieval by agents has been further refined. Additionally, many features have been added that contribute to improving the efficiency of the entire support center, such as customer intent analysis through multimodal support (text + voice + image) and advanced inquiry routing.

Field Service (Dynamics 365 Field Service)
In field service operations, Copilot now provides "Copilot-first experiences." For inspection tasks, it automatically generates checklists from inspection templates, allowing field technicians to perform inspections without omissions.

For schedulers (dispatchers), it provides actionable insights derived from past history, and Copilot advises on which tasks should be prioritized. Seamless integration with Teams and Outlook has also been strengthened so that field staff and managers can quickly access the information they need (such as manuals and inventory status).

Finance and Supply Chain (Dynamics 365 Finance / SCM)
Copilot also contributes to the automation of business flows and the provision of insights in these core operations. In accounting, agent functions are utilized for the automation of tax management and account reconciliation, with Copilot preparing complex expense journal entries and bank reconciliations.

In the supply chain, it automatically generates emails for communication with suppliers, and the AI presents factor analysis (cell-level explainability) and generative demand insights from data to improve the accuracy of demand forecasting. In manufacturing, it also contributes to the sophistication of operations by having the AI compare actual data with planned values to detect process deviations and make improvement suggestions.

Business Central for Small and Medium-sized Businesses
AI agents have also been integrated into Dynamics 365 Business Central, an ERP for SMBs, allowing order processing, report generation, and inventory optimization to be executed via natural language instructions.

Especially in tasks such as ordering and manufacturing instructions, Copilot can now be entrusted with a series of processes like "check the inventory status of XX and automatically place an order if stock is low," enabling even small companies to improve operational efficiency with AI.

Use cases in the business domain
Because Dynamics 365 Copilot is designed to meet the needs of various industries and job functions, it is used in ways that are directly linked to specific business scenarios.

Sales support
In sales teams, Copilot automatically summarizes meeting records and suggests next actions, helping representatives prevent missed follow-ups. For example, along with a summary stating, "Price negotiation is a task from last week's meeting with the customer," Copilot might suggest, "Would you like to start the discount approval process?" It is also useful for knowledge sharing, such as when new sales staff immediately retrieve and reference information from similar past cases using Copilot.

Customer service
In call centers, while an operator is handling a call, Copilot transcribes and analyzes the customer's statements in real-time, displaying appropriate answer candidates and related FAQs on the screen. Operators not only use these suggestions to respond, but after the call ends, Copilot summarizes the interaction history and automatically records notes for the next person handling the account. For supervisors, Copilot also provides a feature that analyzes trends and problem patterns from large volumes of customer feedback, which is then used for service improvement.

Inventory and logistics management
In the supply chain field, Copilot performs demand forecasting and inventory analysis, warning that "Product A is expected to be out of stock next month" or automatically creating order proposals.

Staff only need to review and approve the purchase orders generated by Copilot, reducing the risk of stockouts due to overlooked data. In logistics, AI proactively identifies orders likely to experience delivery delays and automatically drafts notification emails to customers, enabling proactive responses.

Evaluation by adopting companies
Dynamics 365 Copilot is increasingly being positioned as the core of companies' digital transformation strategies. Especially in sales and service areas, there are comments such as, "Thanks to Copilot, data entry work into CRM systems has been reduced, allowing us to focus on actual customer service."

Additionally, in field service operations, field staff have evaluated it by saying, "Searching for necessary information has become significantly faster, making on-site responses smoother." In the finance department, Copilot contributes to expense report creation and monthly closing preparations, and it is well-received by management, who say, "The time available for analysis has increased."

Latest updates for Microsoft Copilot

Microsoft Copilot is at the forefront of rapidly evolving AI technology, and its capabilities are expanded and improved through continuous updates. This section examines major recently announced updates, notable features on the Microsoft 365 roadmap, and the future direction of Copilot's development.

Recent major updates

Microsoft frequently adds new features and enhances existing ones across the entire Copilot family, aiming to improve user productivity and the AI experience. Recent major updates, centered around the May 2025 announcement, include the following:

These updates demonstrate that Copilot is evolving to be more intelligent, support a wider range of tasks, and meet the security and governance requirements demanded by enterprises.

Notable Features on the Roadmap

　The Microsoft 365 public roadmap provides information regarding the preview availability and general availability dates for new features, including those related to Copilot. Since it is updated continuously, regular checks are recommended to stay informed with the latest information.

　Of particular note, the policy to control the visibility of the Copilot Chat icon in Edge for Business (EdgeMicrosoft365CopilotChatIconEnabled), mentioned earlier, is listed on the roadmap with a rollout scheduled to begin around August 2025. The enhancement of such management features is a crucial step in enabling organizations to finely tune their use of Copilot in accordance with organizational policies.

Conclusion

　Microsoft Copilot is a powerful and multifunctional suite of AI assistants capable of delivering value across a wide range of areas, from improving individual productivity to driving corporate digital transformation.

　By leveraging OpenAI's advanced large language models and integrating with Microsoft Graph (especially in Microsoft 365 Copilot) to deeply understand the user's work context, it provides personalized support that sets it apart from conventional AI assistants.

　Its diverse editions and deep integration into essential daily applications such as Word, Excel, PowerPoint, Outlook, and Teams demonstrate the flexibility to meet the needs of various users and organizations.

　However, to ensure successful adoption and utilization, several key factors must be considered.

Success Factors for Adoption and Utilization

1. Setting Clear Objectives
   　It is essential to clearly define the purpose of adopting Copilot in advance (e.g., streamlining specific business processes, reducing time spent on routine tasks, enhancing data analysis capabilities, promoting innovation, etc.). Without a clear objective, it becomes difficult to measure the effectiveness of the adoption.

2. Selecting the Appropriate Plan
   　Because the Copilot family is extensive, it is necessary to select the optimal edition and licensing plan by comprehensively considering the scale of the individual, team, or entire organization, specific needs, desired features, and budget.

3. Phased Adoption and Evaluation
   　Rather than a company-wide rollout, a small-start approach (such as conducting a pilot program) targeting specific departments or use cases is recommended, gradually expanding the scope while objectively measuring and evaluating the results.

4. Employee Education and Skill Development
   　To maximize Copilot's capabilities, appropriate education and training for employees are essential. It is necessary to acquire skills such as how to create effective prompts (prompt engineering), how to critically evaluate information generated by AI (AI literacy), and knowledge regarding the ethical use of AI.

5. Strong Governance and Security Framework
   　Especially for corporate use, it is extremely important to establish clear policies for data protection, access management, and compliance, and to develop the technical and organizational structures to implement them.

　We have introduced Copilot, covering everything from its features and integration to pricing and use cases.

　For business users, there is the benefit of streamlining document creation and meeting responses, allowing more time to be spent on creative work. For engineers, it will speed up coding and debugging, and lower the barrier to learning new technologies.

　For marketing personnel, it can resolve bottlenecks in content production, and for sales and support staff, it can lead to better knowledge utilization and improved response quality, among various other benefits depending on the job role.

　On the other hand, when selecting tools, perspectives such as "does it have the integration that fits our company's usage environment," "can it safely handle the data that needs to be protected," and "is it cost-effective" are important.

　For example, for companies centered on Microsoft 365, M365 Copilot will have high affinity; for those primarily using Google, Duet AI will be a good fit; and for development organizations, GitHub Copilot will provide direct value.

　In information departments, Slack/Zoom AI can improve communication efficiency, while Jasper and Notion AI can be powerful for creative roles. While it is certainly possible to introduce multiple tools, each comes with learning costs and governance requirements, so it is desirable for IT departments and business units to cooperate in drawing up a roadmap.

　Overall, as of 2025, it can be said that generative AI is transforming the way office work is done. Overseas, it is even said that "AI literacy is the new PC skill," and domestically, its active use is being promoted in the context of DX advancement. Moving forward, as competition among companies and technological progress continue to improve accuracy, functionality, and pricing, an era will arrive where more business users and engineers use AI assistants as a matter of course. To avoid being left behind by this wave, it is important to identify the tools suitable for your company and work toward organizational adoption by starting small.

Afterword

　In this article, we have examined the possibilities brought about by Copilot, which is deeply integrated into Microsoft 365, from various angles. The trend of integrating generative AI with office applications provides not only efficiency but also opportunities for new creation. However, user literacy and the establishment of information management are essential. While more enhanced features will surely appear as technology evolves, whether or not you enjoy those benefits will depend on your efforts.

　I hope that the ideas and knowledge gained through this article will be of some help to your business. If you felt that this article was helpful, it would be encouraging if you could "like" or "follow" it. I will continue to share practical know-how and the latest AI trends, so I would be happy if you continue to read my work.

⚠️ Important: Less than 30% of people can fully utilize ChatGPT
Why? It is the difference in "questioning ability."
In fact, there is a scientific law to "how to frame questions" that extracts 10 times the value from AI. It is a method proven by Google's 20% rule and recommended by Amazon's Bezos. Everything is revealed in "The Strongest Skill in the AI Era: Curiosity Power."
Now that AI can return "answers" in an instant, the deciding factor lies in the "question." Please also check out my book, "The Strongest Skill in the AI Era: Curiosity Power," which condenses that secret and 13 practical tools you can use in your work tomorrow.