Claude and Copilot both became one app this month. Which one is your team actually on?

Four things. On 16 September Anthropic merged Claude chat and Claude Cowork into a single app, and added Claude Docs and Claude Slides alongside Claude Design inside conversations, all in beta on paid plans with Enterprise admins choosing when to enable them. On 22 September it launched Claude Opus 5.5, which Anthropic says performs at the level of Claude Fable 5.1 on most work and costs 40% less to run than Opus 5. On 23 September it opened the Claude Marketplace with more than 2,000 connectors and plugins. And on 15 September it expanded Claude for Small Business to 43 workflows and 27 integrations including Xero, Shopify, Stripe and Zapier.

On 25 September Microsoft unveiled a single Copilot app with three sections: Home, which combines chat with the Cowork agent; Code, which lets people build apps and tools by describing them; and Autopilot, always-on agents that run inside a company's Microsoft 365 environment with their own identity and email address. Full Word, Excel and PowerPoint now run inside Copilot. Users can pick OpenAI's GPT, Anthropic's Opus, or an Auto mode. Pricing splits into a per-user licence for everyday use and usage-based billing for Cowork, Code, Autopilot and the newest frontier models, off by default with admin spending limits. Most features roll out over the coming weeks; Autopilot is in private preview.

Less than the vendors would like, more than most employers think. The two products now look similar on paper: chat, long-running agents, documents, connectors. The differences that matter for a UK business are where the work runs, how it is billed, how it is governed, and which one your people will actually open. Microsoft's pitch is that Copilot lives inside your Microsoft 365 tenancy with existing security and compliance, and that you can swap models. Anthropic's pitch is the model itself, enterprise controls that keep your data out of training, and a marketplace of tools. Choose on those, not on the demo.

Four questions. Where does your data already live, and which tool can use it without export? What is your governance model for agents that act on your behalf, and which product's admin controls match it? How will you pay, given Copilot's agent features are now usage-billed and Claude's are plan-based? And who in the business is going to become expert in the tool, because a licence nobody is trained on is a cost, not a capability. Answer those and the choice usually makes itself.

Not if they were trained properly. Someone trained on prompting technique, workflow design, automation logic and governance moves between Claude and Copilot in an afternoon, because the underlying skills are the same. Someone trained on where the buttons were in the June interface is stuck. That is why our Level 4 teaches the RICF prompt framework and automation design first, and the vendor tool second, and why the same standard exists in Claude, Copilot and Gemini editions.

Autopilot is Microsoft's name for always-on AI agents that people create and name themselves. Each runs in the cloud inside your Microsoft 365 environment with its own identity, memory and email address, and colleagues can tag it in Teams or Outlook. Microsoft's own Copilot lead described the autonomy that makes such agents useful as the same thing that makes them terrifying to an IT admin. It is in private preview, so nobody needs to panic, but any business planning to use it needs the same four things we recommend for any agent: a written scope, a log, a tested off switch and a named owner.