Insight

Microsoft Copilot Update: Claude AI Is Now Available in Microsoft 365

Microsoft has announced a major update to Microsoft 365 Copilot, introducing Anthropic’s Claude AI models as an additional option alongside existing OpenAI models. This marks a significant shift to a multi‑model Copilot experience, giving organisations more flexibility in how AI supports their everyday work.

What’s Changed?

Until now, Microsoft Copilot was powered primarily by OpenAI models. With this update, Claude Sonnet and Claude Opus models from Anthropic are now available inside Microsoft 365 Copilot for eligible tenants. 

Claude is being introduced across several areas of the Copilot ecosystem, including:

  • Copilot Chat
  • The Researcher agent
  • Copilot Studio (for custom AI agents)

 

This change is part of Microsoft’s broader strategy to offer model choice, allowing customers to select the AI model that best fits a specific task rather than relying on a single provider.

Where Will You See Claude in Copilot?

Copilot Chat

Licensed Microsoft 365 Copilot users can now choose Claude Sonnet directly from the model selector in Copilot Chat (depending on tenant settings and region). This allows users to compare outputs between Claude and existing models for tasks like drafting, summarising, or analysing content.

Researcher Agent

Microsoft’s Researcher agent, designed for complex, multi‑step analysis, can now run on Claude Opus. This is particularly useful for deep research tasks such as market analysis, policy reviews, or long‑form reporting that spans multiple documents and data sources.

Copilot Studio

In Copilot Studio, organisations can build custom AI agents powered by Claude models. This enables advanced reasoning, workflow automation, and agent‑based processes that can be tailored to specific business needs. 

Why Microsoft Is Adding Claude

Microsoft has been clear that this update is about choice and flexibility, not replacing OpenAI. OpenAI models remain a core part of Copilot, but Claude adds a complementary option with different strengths—particularly in structured reasoning and long‑form analysis.

According to Microsoft, this multi‑model approach allows Copilot to:

  • Match the right model to the right task
  • Improve resilience and performance
  • Support more advanced agent‑based workflows across Microsoft 365

 

Security, Privacy, and Compliance Considerations

Microsoft has confirmed that Enterprise Data Protection continues to apply when using Claude within Microsoft 365 Copilot. Anthropic operates as an approved Microsoft sub‑processor under Microsoft’s existing data protection agreements.

However, there are some important nuances:

  • Claude models are hosted outside Microsoft‑managed environments
  • Some regions (including EU, EFTA, UK Government and sovereign clouds) do not have Claude enabled by default
  • In certain tenants, admins must explicitly opt in before users can access Claude

 

When Is This Available?

  • Frontier / early release: Available now for selected tenants
  • General rollout: Completing in late March 2026 for most commercial Microsoft 365 Copilot customers
  • Availability depends on licensing, region, and admin settings

 

What This Means for Your Business

This update represents a meaningful evolution in how Microsoft Copilot works:

  • More flexibility in AI behaviour and output
  • Better support for complex analysis and research
  • Improved options for organisations building custom AI agents
  • No change to existing Copilot licences, just expanded capability

 

For most users, day‑to‑day Copilot usage won’t suddenly change—but those who need deeper reasoning or more specialised AI workflows now have another powerful option to draw from.

If you’d like help understanding whether Claude is enabled in your tenant, or guidance on when to use Claude vs OpenAI models in Copilot, feel free to ask—we’re happy to help.

 

Let's get started!

looking for IT expertise, advice, or support, please get in touch