Microsoft Copilot for nonprofits - should your charity pay for it?

Microsoft Copilot for M365 is available to nonprofits at around £20 per user per month ($25.50, a 15% discount on the commercial price). It puts AI into Word, Excel, PowerPoint, Outlook, and Teams. You need a qualifying M365 plan underneath it, which most charities on Microsoft's nonprofit programme already have.

Separately, the free Microsoft Copilot (the rebranded Bing Chat) is available to everyone and doesn't require any subscription. It's worth trying before you spend anything.

One thing to note early: Copilot is an easy way to get approval from IT or Ops because it's "already Microsoft." But don't assume that means the data processing is materially different from other AI tools. In our non-legal reading, Copilot's data processing terms aren't meaningfully different from what you'd get with Claude or OpenAI. Look at the small print before using "it's all in Microsoft" as a reason to skip governance questions.

What's potentially useful

Teams meeting summaries are the feature most likely to justify the cost. If your charity runs a lot of meetings, Copilot can produce structured summaries with action items. In our testing, these work reasonably well and can save 15-20 minutes of note-taking per meeting. For organisations that lose decisions in email chains and forgotten minutes, this has value.

Outlook email drafting and summarisation works too. Copilot can draft replies, summarise long email threads, and help manage an overwhelming inbox. The quality varies, but for routine correspondence it saves time. Word drafting assistance is decent for first passes: ask it to draft a report section or restructure a document and you'll get a reasonable starting point. It's not going to write your impact report for you, but it can give you something to edit rather than a blank page.

Where it falls short

Excel analysis was the feature we were most excited about for charity use cases, analysing donor data, programme outcomes, financial patterns, and it's the most disappointing. Copilot in Excel struggles with messy data (missing cells, inconsistent formatting, merged cells), which describes most charity spreadsheets. When it works, it's impressive. It doesn't work reliably enough to recommend as a primary reason to buy.

PowerPoint generation produces generic, corporate-looking results. You'll spend as much time fixing them as you would have spent building from scratch. And across all features, Copilot has no idea how your charity talks, who your beneficiaries are, or what the sector cares about. Everything it produces sounds like a generic nonprofit until you edit it.

There's a broader limitation. Copilot is fundamentally a personal productivity tool. It helps individual users work faster within their own applications. But the harder problems charities face aren't personal productivity problems. They're organisational ones: processing thousands of feedback forms, integrating data across systems that don't talk to each other, building workflows that connect what happens in a CRM with what happens in programme delivery. Copilot doesn't break silos. It can't work with data outside the Microsoft ecosystem. It reinforces existing structures and permissions rather than enabling new ways of working. For the problems that really move the needle, charities need solutions designed around their specific challenges, not generic AI bolted onto existing software.

The cost question

At around £20 per user per month, the maths gets challenging quickly. A charity with 20 M365 users would pay roughly £4,800/year. For 50 users, that's around £12,000/year. These aren't small numbers for most charities.

The question isn't whether Copilot is useful. It can be. The question is whether it solves a problem you actually have, or whether you're accepting the push from Microsoft because the button is there. Free tools like ChatGPT, Claude, and Google's Gemini can do much of what Copilot does in Word and Outlook, just not from inside the application. If your team is willing to copy and paste between tools, the free alternatives cover most use cases.

Where Copilot might justify its cost is in Teams meeting summaries (no free alternative does this inside Teams) and in organisations where staff won't adopt a separate AI tool but will use features built into software they already use.

If you're going to try it, start with 3-5 licences for the people who'll use it most: a fundraiser who writes a lot, a programme manager who's in meetings all day, someone who processes a lot of email. See whether they actually use it after a month. Don't buy it for everyone upfront. £4,800/year buys a lot of ChatGPT Plus subscriptions at £20/month per user, and your team might get more value from a few paid subscriptions to whichever tool they find most useful.

The danger for charities is the per-user pricing model. AI tools that charge per seat become expensive in organisations with lots of users, and charities tend to have lots of M365 licences because Microsoft's nonprofit programme makes the base product affordable. Adding around £20/user/month on top changes the economics significantly. Keep an eye on the pricing; Microsoft has a track record of adjusting nonprofit rates as products mature.

Copilot is a reasonable personal productivity tool. Teams meeting summaries are potentially the most useful feature for charities. But if someone in your organisation is hoping Copilot will solve the bigger challenges (data integration, process automation, bespoke analysis), they'll be disappointed. Those problems need something built for purpose, and buying 50 Copilot licences won't get you closer to solving them.