Anthropic Models in Microsoft 365 Copilot: Enterprise-Grade Compliance and Model Choice
Starting January 7, 2026, Anthropic models will be covered under Microsoft’s enterprise-grade compliance and data handling commitments. As part of this change, access to Anthropic models in Microsoft 365 Copilot and Copilot Studio will be on by default for most customers in the commercial cloud. Anthropic models are delivered with the same enterprise-grade security, privacy, and compliance commitments applied across Microsoft services.
Where Anthropic Models Are Available
Anthropic’s Claude models are integrated across multiple Copilot experiences:
- Microsoft 365 Copilot
- Researcher agent: Users with a Microsoft 365 Copilot license can choose Anthropic models for deep research tasks.
- Agent Mode in Excel (Frontier): Users can select Anthropic models in preview through the Frontier program.
- Word, Excel, PowerPoint Agents (Frontier): Available in Copilot Chat for creating high-quality Office artifacts, also in preview through the Frontier program.
- Copilot Studio: Makers can select Anthropic models when building custom agents, enabling tailored solutions for unique business needs.
Key Benefits
- Enterprise Data Protection: Enterprise-grade compliance and data handling commitments are in place, as shown by the ‘green shield’ indicator in Microsoft 365 Copilot.
- Simplified Experience: Tenant admins no longer need to opt in to Anthropic’s separate commercial terms.
- Model Choice with Confidence: You can build custom agents in Copilot Studio and leverage Anthropic models in Agent Mode in Excel and Researcher without extra configuration or loss of trust.
This flexibility and compliance alignment make Anthropic models a powerful addition to the Microsoft 365 Copilot ecosystem, giving organizations more options while maintaining enterprise-grade security and privacy.
To learn more click here.