Copilot APIs for M365 Developers: 01 - Getting Started — A Microsoft 365, SharePoint & Copilot Blog

Hi everyone!

Today I want to talk about something that’s been on my radar for a while and that I’m truly excited about: the Microsoft 365 Copilot APIs. These APIs are now available on Microsoft Graph and they open up incredible possibilities for us M365 developers to integrate AI-powered capabilities directly into our custom applications.

What are the Copilot APIs?

The Microsoft 365 Copilot APIs enable you to securely access Microsoft 365 Copilot capabilities in your own applications and custom engine agents, while aligning with the compliance standards of Microsoft 365.

Enterprise developers face a critical challenge today: how to harness the power of AI while maintaining secure access to organizational knowledge, without compromising compliance or governance. Traditional approaches force you to build complex AI systems from scratch and extract sensitive data into external sources. The Copilot APIs offer access to production-ready AI capabilities that work directly with your Microsoft 365 data.

The Complete Suite of Copilot APIs

Microsoft has released a comprehensive set of APIs that cover different scenarios. Let me walk you through each one.

⚠️ Important Note on API Status: At the time of writing, some of these APIs are in Preview status. Understanding API versioning in Microsoft Graph is crucial:
v1.0 endpoint (graph.microsoft.com/v1.0/copilot): Stable, production-ready APIs with full support. These are Generally Available (GA).
Beta endpoint (graph.microsoft.com/beta/copilot): Preview version of the API for testing new features. Beta APIs are subject to change without notice, may have slower response times, and are not recommended for production use. Features might be modified or removed, potentially causing breaking changes to your app.
Preview APIs: Some APIs are explicitly marked as “Preview”. These are early-stage capabilities that Microsoft is testing. They give us a chance to experiment and provide feedback, but should be used with caution.
I’ll indicate the status for each API below. Always check the official documentation for the most current status.

🔍 Retrieval API (GA)

Retrieve relevant information from Microsoft 365 content in a secure and compliant way. This API allows you to connect your own AI models to Microsoft 365 content without data extraction. Perfect for building specialized assistants that respond based on your organization’s up-to-date documentation, policies, and knowledge bases.

🔎 Search API (Preview)

Perform hybrid search (semantic and lexical) across OneDrive, SharePoint, and SharePoint Embedded content using natural language queries. Build AI-powered search applications that help users discover relevant documents using natural language rather than exact keyword matches.

💬 Chat API (Preview)

Enable conversational experiences powered by Microsoft 365 Copilot in your custom applications. Integrate Microsoft 365 Copilot into your enterprise applications to answer questions, perform tasks, and provide guidance based on Microsoft 365 data and user context.

💡 My take: The Chat API is the one that excites me the most. The ability to embed Copilot conversational capabilities directly into custom applications opens up incredible possibilities. I see huge potential here for building truly intelligent enterprise solutions.

📅 Meeting Insights API (GA)

Extract AI-generated meeting notes, action items, and discussion topics for Teams meetings. Create applications that automatically extract and organize action items, decisions, and summaries from Teams meetings and link them with project management tools or CRM systems.

📊 Interaction Export API (GA)

Enable compliance solutions to capture and archive user interactions with Copilot across Microsoft 365 applications. Essential for regulated industries where documenting AI-assisted content generation is necessary, but that’s not all. You can also use this API to monitor usage patterns, measure adoption, analyze the value Copilot brings to your organization, and much more. Honestly, the only limit here is your imagination!

💡 My take: These APIs are incredibly powerful, among the most interesting in my opinion. However, they should be used with caution. Since they expose all user prompts and interactions, there’s a potential privacy concern if not implemented properly. Fortunately, Microsoft has designed them with specific permissions that ensure secure usage when configured correctly. Definitely worth exploring!

🔔 Change Notifications API (Preview)

Subscribe to change notifications for Copilot interactions across Microsoft 365. Build applications that monitor and log AI interactions in real time, enabling proactive compliance checks and auditing.

📈 Copilot Usage Reports API (GA)

Query user counts and usage data for Microsoft 365 Copilot in your organization. Build reports to track the adoption of Microsoft 365 Copilot by your users.

📦 Package Management API (GA)

View and manage apps and agents across Microsoft 365. Create an inventory of all agents and apps within your organization.

💡 My take: Finally, the Swiss Army knife for managing agents, or at least the beginning of it! This API is a great starting point for organizations looking to get a grip on their growing ecosystem of agents and apps.

Key Benefits

🔒 Secure Grounding, Governance and Compliance

Access Microsoft 365’s knowledge index directly. All existing permissions, sensitivity labels, compliance controls, audit, logging, monitoring, and policy enforcement are automatically respected.

🚀 Production-Ready AI

Accelerate the development of your custom RAG pipelines and meeting transcription processing by using the same production-grade AI capabilities that power Microsoft 365 Copilot. Whether you develop your solutions using Azure AI Foundry, Microsoft Copilot Studio, or the Microsoft Agents SDK, the Copilot APIs are seamlessly integrated.

🛡️ Responsible AI

When you use these Copilot APIs, you’re protected against harmful content via RAI validation checks.

Copilot APIs vs. Microsoft Graph APIs

This is an important distinction to understand:

Microsoft Graph APIs provide CRUD operations on Microsoft 365 data
Copilot APIs deliver AI-powered capabilities built on Microsoft 365 data

Use Microsoft Graph APIs when you need to manipulate and access data; use the Copilot APIs when you need AI to reason over that data. This is also reflected in the licensing model: Graph APIs are available under standard Microsoft 365 license terms, while Copilot APIs require a Microsoft 365 Copilot license.

Requirements

Before getting started, make sure you have:

Microsoft 365 Copilot License - Required for each user who accesses Microsoft 365 Copilot functionality via these APIs
Microsoft 365 Subscription - An E3 or E5 subscription (or equivalent) as the foundation
App registration in Microsoft Entra ID - With appropriate permissions

The Copilot APIs are available as standard REST APIs under the Microsoft Graph namespace (graph.microsoft.com/v1.0/copilot and graph.microsoft.com/beta/copilot). They use the same authentication and authorization process as other Microsoft Graph APIs.

📋 Terms of Use: Using the Microsoft 365 Copilot APIs is subject to the Microsoft 365 Copilot APIs Terms of Use (preview). I cannot stress this enough: read them carefully before you start building. These terms cover important aspects like acceptable use, data handling, and compliance requirements. Ignoring them could put your project or your organization at risk.

Why I’m Excited: SPFx Integration

In my opinion, the true power of these APIs really shines when you combine them with SharePoint Framework (SPFx). Think about it:

Build web parts that leverage AI-powered search across your organization’s content
Create custom experiences that extract meeting insights directly in SharePoint pages
Develop solutions that provide contextual AI assistance within your intranet
Integrate Copilot conversational capabilities into your custom dashboards

The combination of SPFx’s deep integration with SharePoint and Microsoft 365, together with the Copilot APIs’ AI capabilities, opens up scenarios that were simply not possible before.

I’m already working on some sample projects that demonstrate how to use these APIs within SPFx solutions. Stay tuned, I’ll be sharing them soon!

This is Part of a Series

This is the first post in my series about Microsoft 365 Copilot APIs. Here’s what’s coming next:

01 - Getting Started (this post) - Overview of all available APIs
Coming Soon - Deep Dive into the Chat API
Coming Soon - Building Search Experiences with the Search API
Coming Soon - Exporting Copilot Interactions for Compliance
Coming Soon - SPFx Web Parts powered by Copilot APIs

I’ll update this list with links as new posts are published. Subscribe to the RSS feed or check back regularly!

Useful Resources

To dive deeper, here are links to the official documentation:

Wrapping Up

The Microsoft 365 Copilot APIs represent a significant step forward for us M365 developers. Instead of building complex AI systems from scratch, we now have access to production-ready, secure, and compliant AI capabilities that work directly with our Microsoft 365 data.

Whether you’re looking to build intelligent search experiences, extract insights from meetings, create conversational interfaces, or ensure compliance with AI interactions, there’s an API for that.

I’m genuinely excited about what we can build with these tools. As I mentioned, I’m working on some SPFx samples that will demonstrate practical implementations—keep an eye on this blog for updates!

What scenarios are you most excited about? Let me know in the comments!

See you soon! 🚀

Federico

Content Creation Note: Portions of this post were prepared with assistance from AI tools including Copilot for M365 and GitHub Copilot, to streamline the writing process. All content has been reviewed and edited to ensure accuracy and quality.

Copilot APIs for M365 Developers: 01 - Getting Started