AI strategy for charities is still rare, even though adoption is not. Eight in ten charities already use AI in some form. That gap sat at the centre of a recent Charity Recruit boardroom led by adviser Darren Richards.
“It’s like when you order a meal and it comes with fries on the side. You didn’t ask for the fries, they just came and they’re not helping your waistline.”
Fundraiser Turned AI Adviser
Darren Richards founded Charity AI Partners and is known online as the Charity AI Guy. He used that image to describe how AI turns up inside everyday charity software. It arrives bundled into a CRM or an email platform, chosen by no one in particular. Most charities end up using it long before anyone agrees what it should never do. Darren was named in the AI 100 UK list this year and he has also written a book on charity leadership called The Plain Parable.
Richards spent 24 years working as a fundraiser before moving into AI advice. He pointed to one children’s hospital charity that once saved his own twin sons. That charity used predictive AI on its own mailing data. Its return on one appeal rose from two to one to six to one and it reached that result while mailing half as many supporters as before. The average gift also climbed by almost a third.
Why Most Charities Are Using AI Without a Plan
Results like that remain rare across the sector Richards now advises. Many charities already use AI without knowing how widely, which he called “shadow AI”. Only one in five have a genuine strategy rather than a policy. Most plans simply list which tools staff may or may not use. Fundraising, where the hospital found its gains, rarely gets a mention.
Three Types of AI Charities Need to Know
He grouped everything charities are likely to meet into three broad types. Each type does a different job inside a charity’s fundraising work.
Predictive AI studies past donor data to forecast future giving and legacy gifts
Generative AI produces text and images by predicting the next likely word or pixel
Agentic AI takes independent action toward a goal, running tasks such as research or report writing
Together these three types make up a workable AI strategy for charities.
The Search Shift Charities Cannot Ignore
Richards then turned to how people now find charities online. More than half of web traffic today comes from AI systems rather than people. Seven in ten visitors never click through to a charity’s own site, they read a short summary instead and act on that alone. Richards said charities now need to be visible inside AI answers, not only on page one of Google.
“AI can read billions of words, but it can’t read the room.”
Guardrails That Keep Humans in Charge
Before any charity scales its AI use, Richards set out clear limits. No message should reach a bereaved supporter without a person checking it first and no image or story should be invented without saying so. A risk assessment, often called a DPIA, should also sit behind every use case. Darren warned against phone agents that pretend to be human callers, as accountability always sits with people, never with the tool itself.
In Summary
One point stood out, which was buying software is not the same as having an AI strategy for charities. Richards showed how predictive, generative and agentic AI each play a different role. He also stressed that guardrails matter before any tool goes live. Charities that set governance first, then choose their tools, tend to see stronger returns and the risk drops once that order is followed.
Emily Formby is Divisional Director at Charity Recruit, leading executive and senior recruitment for charities and not-for-profit organisations across the UK. She partners with charity leaders and trustees to secure senior talent, hosts The Charity Champions Podcast, and in 2025 served as a judge at the Third Sector Business Charity Awards.
