Claude Comes to Microsoft 365: What This Means for UK Businesses in 2026

When Claude Opus 5 dropped in late July, most of the commentary landed on the model improvements, and they’re genuinely significant. Opus 5 is Anthropic’s most capable model yet, sitting at the top of the benchmark charts by a comfortable margin.

That wasn’t the story I was watching.

The story was this: Claude can now take action inside Microsoft 365.

Not chat. Not answer questions from a sidebar. Actual action: booking meetings, drafting and sending emails, working across Word, Excel, Outlook, and Teams natively, with nothing to copy or paste.

For UK SMEs already paying for Microsoft 365, that changes the AI equation in a way I don’t think the industry has fully absorbed yet.

What Claude in M365 actually does

Worth being specific here, because the distinction carries the whole story.

Earlier AI integrations in Microsoft 365, including a lot of the early Copilot experience, would surface information and generate content for you to then act on yourself. Copilot would draft an email; you’d copy it into Outlook and send it. Copilot would summarise a document; you’d read the summary in a side panel and take it from there.

Claude’s M365 integration works differently. Prompt it with something like “set up a meeting with Lisa for 10am Tuesday” and Claude goes and does it: checks Outlook availability, creates the meeting, adds the attendee, sends the invite.

That sounds like a small thing. It’s the gap between AI as a content generator and AI as something that actually acts on your behalf, and it’s the shift we’ve been building toward for two years.

The Copilot comparison

This lands harder set against what Microsoft Copilot has been doing lately.

Copilot Cowork, Microsoft’s agentic AI offering, launched on a consumption billing model. Rather than a fixed monthly subscription, businesses pay per task using “Copilot Credits.” Nobody can tell you what a Copilot Credit is actually worth in real-world terms, because Microsoft has never published a conversion rate from tokens to credits.

We ran a direct comparison. Same task, same model (Opus 4.8), same brief. The output was near-identical. The invoices weren’t.

Claude completed the task for $0.92. Copilot Cowork: $2.22.

More than double, for the same output.

Now add that Claude Cowork, Anthropic’s equivalent offering, has been available as a standalone product since August 3rd, 2026. No Microsoft subscription required. With native M365 integration now live, Claude can do what Copilot does, inside the same tools your team already uses, at a fraction of the cost.

Why this is a strategic moment for UK SMEs

I’ve been advising UK SMEs on AI adoption for two years now. For most of that time my default advice was simple: if you’re already on Microsoft 365, Copilot is your most natural entry point. Same interface, familiar tools, easy procurement.

That advice has got more complicated.

The reasons to consider Claude alongside, or instead of, Copilot have stacked up fast.

Cost. The Copilot Credits model is opaque and, based on our testing, expensive relative to Claude for equivalent workloads.

Capability. Opus 5 is currently the most capable commercially available model. On complex tasks, strategy documents, multi-step research, detailed analysis, the quality gap against Copilot-native models is real.

Integration. With M365 integration live, the old objection (“but we’re already in the Microsoft ecosystem”) has lost most of its force. Claude now works where your team already works.

Transparency. Anthropic publishes model capabilities, pricing, and limitations clearly. When a board asks what you’re paying and why, that question is far easier to answer with Claude than with Copilot Credits.

What you should do now

I’m not telling every UK SME to cancel Copilot tomorrow. The right call depends on your workflow, your team size, and what you’re actually using AI for today.

What I am saying is that any business making AI procurement decisions right now should be running an evaluation, not assuming one tool is the answer by default.

Here’s the framework we use:

1. Identify your top three AI use cases, the tasks your team would reach for AI on most often.

2. Run the same brief through both tools, in real working conditions, not a demo environment.

3. Compare output quality honestly: not “does it work” but “does this actually save us time?”

4. Compare cost at realistic usage volumes, not the headline licence price.

5. Factor in governance: where does your data go, and what does that mean for a UK business specifically?

The AI landscape has moved fast enough in 2026 that defaulting to Microsoft is no longer automatically the lowest-risk or lowest-cost option.

The bigger picture

Anthropic integrating Claude into M365 is a bigger deal than it looks on the surface. It challenges the assumption that owning the productivity suite means owning the AI layer sitting on top of it.

For UK businesses, that’s good news. Competition pushes capability up and cost down, and having genuinely strong alternatives integrated into the tools you already use hands you more leverage than most people realise.

The real question was never which AI to use. It’s whether you’ve actually tested the options properly.

Most businesses haven’t. That’s where the opportunity sits.

Matt Neal is the founder of Artificia1. We help UK SMEs navigate AI adoption practically, from tool selection and procurement to team training and strategy. Talk to us if you’re evaluating your AI options.