Anthropic Claude in Copilot for Microsoft 365
Updated: January 2026
Microsoft has officially integrated Anthropic Claude models—including Claude 4.5 Sonnet and Claude 4.5 Opus—into the Copilot for Microsoft 365 ecosystem. This update transforms Copilot into a multi-model platform, giving you the flexibility to choose the best AI engine for your specific tasks
Anthropic-Powered Features:
- Researcher Agent: Tackles complex, multi-step research by reasoning across your emails, files, and the web. You can now toggle on Claude to compare deep reasoning results.
- Copilot Studio: Custom agent creators can now select Claude models to power specialized enterprise agents for automation and deep analysis.
Enterprise-Grade Security & Privacy:
Anthropic has been onboarded as an official Microsoft subprocessor. This means your data remains protected under the Microsoft Product Terms and Data Protection Addendum (DPA).
- Secure Environment: Your data is processed within the same secure framework used for all Microsoft 365 services.
- No Data Training: Your organizational data is not used to train Anthropic's underlying models.
Key Dates & Availability:
- As of January 2026, Anthropic models became "on" by default for most commercial customers. Microsoft expects full feature availability and global rollout to be complete by the end of February 2026.
How to Access:
- Sign in to the Copilot app* (copilot.microsoft.com) with your University of Miami credentials.
- Select Researcher under Agents in Chat.
- Select Try Claude.
*Note: You must have a Copilot for Microsoft 365 license to use Claude in the Copilot app (desktop and web app only). In addition, Claude is available for use during an active session. Once the session ends or the Copilot app is closed, the system automatically reverts to the default generative AI model.
Ready to see how Claude's reasoning compares to GPT? Check out the new models in Copilot today!
If you have any questions, or if you experience any technical issues, please contact the IT Service Desk.
