Copilot change before 9 December
A default that’s about to change for UK Copilot users
If your business uses Microsoft 365 Copilot, there’s a setting worth checking before 9 December. According to The Register, from that date Microsoft will switch on access to Anthropic’s Claude models for UK Copilot tenants whose admins haven’t already chosen a setting. Admins can turn it off, but doing nothing will count as opting in.
That isn’t necessarily a problem. Many businesses will be happy to give their teams access to another set of Ai models inside Copilot. But it is a decision about where your data is processed, and it should be one you make on purpose rather than by default.
What’s changing
Anthropic works with Microsoft as a subprocessor, which means its Claude models can power parts of Copilot and related tools. Microsoft already switches Claude on by default for most commercial tenants. Until now, the UK, the EU and EFTA countries have been the exceptions, with access set to “No users” unless an admin opts in.
The Register reports that this is changing for the UK. From 9 December, UK tenants without a chosen setting will have Claude access turned on. Microsoft is also retiring a setting that limited Claude to Copilot in the Microsoft 365 apps only. Microsoft says the move “affirms Microsoft’s commitment to offering choice between leading AI models”.
Why it matters for your data
Microsoft’s documentation says Claude models in Copilot, Researcher, Copilot Studio, Power Platform and Copilot in the Microsoft 365 apps are currently excluded from the EU Data Boundary and, where they apply, in-country processing commitments. The Register notes that, according to Microsoft, Anthropic processes the relevant data in the United States.
For many businesses that will be fine. For others, especially those with contracts, policies or sector rules about where client data can go, it may not be. Either way, it’s worth checking before the default decides for you.
The change only applies where no choice has been made. If an admin in your business has already set this option, that choice stays as it is. The businesses most likely to be caught out are those where nobody has looked at Copilot’s admin settings since it was first switched on.
What to do before 9 December
1. Check who manages your Copilot settings
Changing this setting needs someone with the AI Administrator or Global Administrator role in Microsoft 365. If your IT is managed for you, ask your provider to confirm the current setting.
2. Find the setting
In the Microsoft 365 admin centre, go to Copilot, then Settings, and open AI providers operating as Microsoft subprocessors. This is where you choose whether Anthropic is available, and to which users or groups.
3. Make a decision and record it
Check your data protection obligations and any client commitments, then choose: allow Claude for everyone, for selected groups only, or not at all. Note the decision and the reason, so it’s easy to show if a client or auditor asks.
4. Tell your team
If you do allow Claude, let people know which models they may see in Copilot and what your Ai usage policy says about client data.
Getting more from Copilot, safely
Settings like this are a good reminder that Copilot isn’t a switch-it-on-and-forget-it tool. Our Microsoft 365 team can review your Copilot setup with you, and our in-person Copilot training on 25 and 26 November covers how to use it well day to day. If you’d like a hand checking this setting before 9 December, speak to Flotek.
Sources: The Register, 10 Nov 2026; Microsoft Learn, Claude models from Anthropic in Microsoft Online Services.
Further reading
Schedule a Free IT Audit & Cost Breakdown
Want it fixed and kept that way? From just £14 a month, we verify your domain and monitor it 24/7, so your email lands and nobody else can send as you.
