A small charity with two paid staff and a rota of volunteers carries much the same paperwork as an organisation ten times its size. Trustee minutes, funder reports, Gift Aid records, safeguarding logs, volunteer rotas, grant applications. AI will not fix underfunding, but it can take real hours out of that load if you use it with care. This guide covers where it helps, where it goes wrong, and what the regulator expects of you.
Are small UK charities already using AI?
The latest numbers come from the Charity Digital Skills Report 2026. Fundraising magazine reports that 79% of organisations now use AI routinely, from a survey of 807 charities where small charities made up 71% of respondents. Charity Digital's breakdown puts small charity use at 74% and large charity use at 92%.
The gap that matters is between the second and last numbers. Most small charities use AI, but far fewer have agreed rules for it.
Which admin tasks should a small charity hand to AI first?
According to Fundraising magazine's report on the 2026 survey, 63% of charities use AI for admin and project management. That makes it the most common use in the sector. Some low risk starting points:
- Trustee minutes. Paste in rough notes with no names beyond officers' roles and ask for minutes in your usual format, with action points listed by owner.
- Policy refreshes. Ask for a plain English version of your volunteer expenses policy, then compare it line by line with the original.
- Funder guidance. Summarise a long funding prospectus into eligibility, deadlines and scoring criteria before anyone commits a day to the application.
- Recurring emails. Draft the standard reply to "can I volunteer?" or "do you take donated furniture?" once, and save it.
Writing each process down as you go helps too. The approach in this guide to documenting business processes with AI works just as well for a charity handover file as for a shop.
Can AI help with grant applications without making them generic?
Grant work is the second big use. Fundraising magazine reports that 45% of charities use AI for grant fundraising. The problem is that a model writes confidently. Ask it to describe your impact and it will produce polished sentences about outcomes you never measured.
A better workflow:
- Paste the funder's question and word limit.
- Paste your real material: attendance figures from your own records, a short anonymised case study, the budget lines.
- Instruct it to use only the facts provided and to flag any gap with [NEEDS EVIDENCE] rather than filling it.
- The project lead edits every paragraph in their own words.
What does the Charity Commission expect from trustees using AI?
In 2024, Civil Society reported on a Charity Commission blog telling charities to consider an internal AI policy. The regulator said trustees remain responsible for decision making and that it is vital this is not delegated to AI. It also warned that generative AI can confidently produce inaccurate, plagiarised, copyright infringing or biased results.
For a small charity, a proportionate response fits on one page and goes to the next trustee meeting:
- Which tools are approved, and on which accounts (charity accounts, not personal ones).
- What must never be pasted in: beneficiary details, safeguarding notes, health information, donor records.
- Which outputs need a named person to check them before use.
- Who reviews the policy, and when.
Only 36% of small charities have an AI policy, according to Fundraising magazine's coverage. Writing one puts you ahead of most of your peers.
How do we keep beneficiary and donor data safe?
Charities hold some of the most sensitive data around: safeguarding concerns, health conditions, immigration status, addresses of people fleeing abuse. UK GDPR applies to all of it whatever tool you use. Practical rules that small teams actually follow:
- Anonymise first. "Client A, aged 70s, lives alone" is enough for drafting a support plan template.
- Use charity accounts. A volunteer's personal chatbot login leaves you no control over the history.
- Read the data settings. Check whether conversations are used for training and whether you can turn that off.
- Log it. If AI is used in a process involving personal data, note it in your data protection records.
This UK AI data protection checklist walks through the questions to ask before you approve a tool.
What stops small charities getting real value from AI?
In the 2026 survey, 56% of charities named limited skills as the biggest barrier to adoption, Fundraising magazine reports. Charity Digital adds that hesitance among colleagues has risen too.
The pattern in a typical small charity is one keen staff member getting good results while everyone else avoids the tools. The fix is cheap:
- Pick one or two approved tools rather than five.
- Keep a shared document of prompts that work, one per recurring task.
- Run a one hour session where each person completes a real task, not a demo.
- Ask volunteers what they would hand over first. They usually know which forms are worst.
Is an AI tool better than hiring a part-time administrator?
A part-time administrator answers the phone, knows the regular donors by name and notices when a volunteer seems unwell. AI does none of that. What it does well is the clearing work that eats their week: first drafts, reformatting, triaging a full inbox. This comparison of an AI admin assistant against a virtual assistant sets out the trade-offs, and most of them apply directly to charities.
If email is the biggest drain, look at sorting and prioritising email with AI before anything more ambitious.
How should a small charity get started this month?
A realistic four week plan:
- Week one. Draft the policy using the points above and circulate it to trustees.
- Week two. Set up organisation accounts on one tool and switch off training on your data where the option exists.
- Week three. Run two tasks through it, such as trustee minutes and a funder guidance summary. Time them against the old way.
- Week four. Keep what saved time, drop what did not, and write the working prompts into your shared document.
Small charities do not need a large project to benefit. They need a few agreed habits, a clear line on data, and someone who reads every output before it goes out.
AI for Charities UK — FAQ
How many small charities in the UK actually use AI?
Most of them now. According to Charity Digital's write-up of the Charity Digital Skills Report 2026, 74% of small charities use AI in some form, compared with 92% of large charities. Across the whole sector, Fundraising magazine reports that 79% of organisations now use AI routinely. The report drew 807 responses, and small charities made up 71% of them, so the small charity figures rest on a solid sample. The practical point is that using a chatbot to draft an email is no longer unusual. What separates charities now is whether they use AI deliberately, with a written policy and clear rules about data, or leave individuals to try tools on their own.
Do trustees need to approve AI use in a small charity?
The Charity Commission has not written separate AI rules, but it expects trustees to apply their existing duties. In a 2024 blog reported by Civil Society, the regulator said trustees remain responsible for decision making and that it is vital this is not delegated to AI. It also suggested charities consider an internal AI policy. For a small charity that means a short paper at a trustee meeting: which tools staff and volunteers may use, what data can never go into them, and who checks outputs before anything is sent or published. Minute the decision and review it once a year.
What is the safest first AI task for a small charity?
Start with internal drafting that contains no personal data. Good examples are turning rough meeting notes into minutes, rewriting a policy in plain English, drafting a volunteer role description, or summarising a long funder guidance document. These tasks save real time and the risk is low because a person reads everything before it goes anywhere. The Charity Digital Skills Report 2026, as reported by Fundraising magazine, found 63% of charities use AI for admin and project management, so this is where most of the sector already starts. Keep beneficiary names, case notes, health details and donor records out of general AI tools until you have a policy and a tool with suitable data terms.
Can AI write our grant applications?
It can help, but it should not write them alone. Fundraising magazine's coverage of the Charity Digital Skills Report 2026 says 45% of charities use AI for grant fundraising, mostly for first drafts, restructuring answers to fit word limits, and checking against funder criteria. The weak spot is evidence. A model will produce fluent, confident claims about outcomes your charity never measured, and funders notice generic text. Feed it your own monitoring data, case studies and budget, ask it to draft within those facts only, and have the person who knows the project edit every paragraph. Treat the output as a structured first draft, never a submission.
What stops small charities getting value from AI?
Skills, far more than money. Fundraising magazine reports that 56% of charities in the 2026 Charity Digital Skills Report cite limited skills as the biggest barrier, and that only 36% of small charities have an AI policy. In practice that looks like one enthusiastic staff member using a tool nobody else understands, no shared prompts, and no agreement about data. The fix is modest: agree a policy, pick one or two tools, write down the prompts that work for recurring tasks, and run a one hour session so every staff member and key volunteer uses them the same way.
