AI for charities: doing more with a stretched team

If your charity is trying to do more with a team that is already stretched thin, AI is worth a careful look: most charities now use it, and the safe wins are in fundraising copy, comms and admin. The trap is scale without care, so start small, and start with a policy.

Adoption has surged. Roughly three-quarters of charities now use AI in their day-to-day work, up sharply in a year, because it helps stretched teams get more done.

The opportunity is real, and so are the worries: donor and beneficiary data, and the quality of what goes out in your name. That is why the first move is not a clever tool. It is a simple, written rule for how your people use AI, so the time it saves does not come with a risk you did not mean to take.

How far this has moved

The numbers have jumped, and quickly. The 2025 Charity Digital Skills Report found that 76% of charities now use AI tools in their day-to-day work, up from 61% a year earlier. In the same survey, 50% said they were developing an AI policy, a big jump from 16% the year before. Usage is running well ahead of the rules meant to guide it.

A second UK body sees the same gap. Charity Excellence Framework's Future Charity Report on charities and AI, published in April 2026, found that around two-thirds of charities describe themselves as exploring or experimenting with AI, yet fewer than one in four have approved tools, policies or training in place. Use is high, governance is low, and that is the whole shape of the risk.

For a stretched team, that gap is the thing to close. Not by using AI less, but by putting a light rule around it so the value is real and the worry is contained.

Where AI helps a stretched team

Four jobs are where a charity gets time back with the least risk.

Fundraising copy is the first. Drafting appeal emails, grant-application text, thank-you letters and social posts is exactly the kind of writing a tool does a decent first version of, for a person to shape and sign off. It takes the blank-page load off a small team without handing over the voice, which stays yours.

Comms is the second. Newsletters, event listings, website updates and routine social content all start faster when AI produces a draft to edit rather than something to write from scratch. The tool gets you to a first cut; your team keeps the tone and the truth of it.

Admin is the third, and often the quietest win. Summarising a long document, drafting meeting notes, tidying a spreadsheet or turning rough bullet points into a clear briefing are all pattern jobs that give a few hours back each week.

Research is the fourth. Pulling together background for a funding bid, summarising a policy paper or getting a first orientation on an unfamiliar topic is quicker with a tool, as long as a person checks the facts before they are used. In each of these, AI produces the raw material and a person keeps the judgement. That is the line between charities getting real value and charities quietly taking on risk.

The real worries

Two worries are worth taking seriously, because they are the ones that cost a charity trust.

The first is data. Charities hold sensitive information about donors, beneficiaries and service users, and some of it is deeply personal. That information does not belong in a consumer AI tool open in a browser tab. Personal data about the people you serve is covered by data protection law and by the duty of care you already owe them, and pasting it into a public tool can breach both. The rule is simple: identifiable personal information about donors or beneficiaries stays out of any tool that has not been contracted to protect it. This is the same discipline covered in our note on whether staff can put client data into ChatGPT, and for a charity it matters just as much.

The second is quality, which shades into reputation. AI will produce a fluent draft that reads perfectly and contains a wrong figure, a made-up quote or a claim about your work that is not quite true. Sent out unchecked, that lands in a funder's inbox or on your public feed in your name. The answer is not to avoid the tool, it is to make sure a person reads and approves anything before it goes out, especially anything a funder, regulator or beneficiary will see.

Why a simple policy comes first

The reason a policy comes before a clever tool is the gap in the numbers: most charities now use AI, but far fewer have written down how. A one-page rule closes that gap for very little effort, and it does not need to be a formal governance document. It needs to answer three questions in plain language.

Which tools are approved, so people are not each choosing their own. What must never go into an AI tool, chiefly personal data about donors and beneficiaries. And who checks AI-drafted work before it goes out, so nothing public is sent unread. Our simple AI policy for small firms is a good starting point, and it is the cheapest protection a charity can put in place. Write that first, then pick your tools, and the value comes without the risk you did not intend.

Doing it on a tiny budget

You do not need a budget to start, which is the good news for a stretched team. Free and low-cost tiers of the main tools are enough to prove the value on real work, and several vendors offer nonprofit pricing worth asking about before you pay full price. The constraint is not money, it is focus.

So do not spread AI across everything at once. Pick one job that eats time every week, fundraising copy or meeting notes are common choices, and use AI only there for a fortnight. That keeps the learning contained and the wins visible, and it stops the tool sprawl that turns a good idea into a dozen half-used accounts. Choosing that first job well matters more than the tool, which is why our note on how to choose an AI use case is a useful next read.

What to do next

Write the one-page policy first: the approved tools, the data that never goes in, and who signs off public-facing work. Then pick one recurring task your team does every week that never touches personal data about a beneficiary, such as drafting a newsletter or summarising meeting notes. Run it through a business-grade AI tool for two weeks, with a named person checking every output. Measure the time saved against the corrections needed. That gives you a real, bounded sense of where AI helps your charity, with the data and the quality kept firmly under human control.

If it would help to work out where AI fits in your charity, and where a person must stay in charge, book a call and we will think it through with you.

Common questions

Should a charity use AI?

Most already do, and used carefully it helps a stretched team do more. The safe wins are in drafting: fundraising copy, comms, admin and background research, all for a person to check and approve. The move that makes it safe is writing a one-page policy first, so everyone knows which tools are approved, what data must never go in, and who signs off before anything is sent.

Can charities put donor or beneficiary data into AI tools?

Not into consumer tools. Personal information about donors and beneficiaries is covered by data protection law and by the duty of care you owe the people you serve, so it must not go into a public AI tool. If you need AI to touch personal data, use a tool on a proper contract that sets out how the data is handled and confirms it will not be used to train the vendor's models.

Where does AI help a charity most?

In the writing and the admin: appeal and grant copy, newsletters and social posts, meeting notes, document summaries and first-pass research. These give a small team hours back each week at low risk. The judgement, the tone of voice and the facts stay with a person, who checks anything before it goes out in the charity's name.

---

This is general information, not legal advice. Personal data about donors and beneficiaries is covered by data protection law, so take advice from a qualified professional on your own obligations before using AI with personal information.