Building on Microsoft 365 Copilot? Here’s your playbook.

Declarative agents are quickly becoming one of the most exciting ways to extend Microsoft 365 Copilot and bring organizational knowledge, workflows, and tools directly into the flow of work. But as agent capabilities grow, so does the need for practical guidance: How do you build agents that are useful, grounded, extensible, and measurable?

That’s why we’re launching The Microsoft 365 Copilot Agent’s Playbook, a four-part livestream series designed to help developers, makers, and technical teams understand how modern Microsoft 365 Copilot declarative agents are built, extended, grounded, and evaluated.

Across the series, engineers and product makers from Microsoft will walk through real-world patterns, live demos, and practical techniques you can apply as you build with Microsoft 365 Copilot. Each session includes a focused presentation, demos, and live Q&A so you can learn directly from the teams behind the platform.

Register: The Microsoft 365 Copilot Agent’s Playbook | Microsoft Reactor

Who Should Join?

Whether you’re just getting started with Microsoft 365 Copilot extensibility or already building declarative agents, this series is designed for you.

Join us if you’re a developer, maker, architect, or technical decision-maker who wants to learn how to:

• Extend agents with skills and actions
• Ground agent responses with enterprise context using WorkIQ
• Connect MCP apps to create richer interactive experiences
• Evaluate and improve agent quality with Microsoft 365 Copilot Evals

If you’ve been looking for a practical path from “what can agents do?” to “how do I build and improve one?”, this series is for you.

Schedule

We have an exciting lineup of sessions planned, each focused on a key part of the Microsoft 365 Copilot agent development journey. All sessions run from 9:00 AM–10:00 AM PT.

What You’ll Learn

Throughout the series, you’ll build a practical understanding of the Microsoft 365 Copilot agent ecosystem and how the pieces fit together.

You’ll learn how to design agents that do more than respond to prompts — agents that can use skills, reason over relevant context, connect to tools, and improve over time through evaluation.

By the end of the series, you’ll have a clearer playbook for building agents that are:

• Useful: designed around real user tasks and workflows
• Grounded: informed by relevant enterprise context
• Extensible: connected to actions, skills, and MCP-powered experiences
• Measurable: evaluated and improved with modern developer tools

Get Ready for the Series

To get the most out of the livestreams, we recommend reviewing the basics of Microsoft 365 Copilot extensibility and declarative agents before the first session.

You may want to have the following ready:

• Familiarity with declarative agents, and Microsoft 365 Copilot extensibility concepts covered here https://microsoft.github.io/copilot-camp/pages/extend-m365-copilot/ in Copilot Developer Camp
• Questions for the live Q&A with Microsoft product and engineering teams around these topics.

Don’t Miss the Microsoft 365 Copilot Agent’s Playbook

The future of AI agents is practical, contextual, and extensible — and this series is your chance to learn how to build for it.

Join us from August 18 through September 8, 9AM PT for The Microsoft 365 Copilot Agent’s Playbook, and learn how to build, ground, extend, and evaluate agents for Microsoft 365 Copilot.