Claude Is Coming to Copilot: The Integration We’ve Been Waiting For

It was only ever a matter of time...

Anthropic's models are going to be available for use with select Microsoft 365 services from January under a subprocessor arrangement that assures the existing standards of the data guarantees we are used to for M365 Copilot (though if you're in the UK or Europe, you're going to have to wait a bit longer).

As many are eagerly awaiting the arrival of GPT-5.2 as OpenAI's emergency "code red" response to Gemini 3, those with actual work to do are increasingly impressed by Claude Opus 4.5. Certainly, I now have multiple clients that tend to rely on Anthropic's models for a variety of reasons that mean we should all be excited for their broader arrival into the Copilot ecosystem.

The situation created a couple of months back when a Claude x Copilot collab was first announced was never sustainable in the long term. "I think Claude is great, but you should definitely not turn this on if you care about the reasons I've told you to rely on Microsoft's governance and security credentials", is not a comfortable position as a consultant.

Ultimately, Copilot services will almost certainly be better with access to Claude for at least some features. But from my perspective, I would have happily waited until the right deal with Anthropic was inked to avoid the confusion of a shiny new feature piercing the data wall Microsoft traditionally has put around customer tenants.

I can't wait until January, when I'll finally get to turn this on!

First posted on Linkedin on 12/15/2025 -> View Linkedin Post Here