79% of UK charities use AI. Only 4% use it strategically.
Seventy-nine per cent. That’s the number leading many headlines about the Charity Digital Skills Report 2026. It’s the ninth annual survey of the UK charity sector, and the biggest yet, with a record 807 responses. Coverage of the launch highlights how nearly four in five UK charities now routinely use AI, up from 61% just two years ago.
It’s a really fast adoption curve, but it’s also close to meaningless on its own.
Page 13 of the report gives a clearer picture. Only 4% of charities say AI is 'widely used and part of our strategy,' up just two points from last year. Another 34% use AI in some areas, but not as part of a plan. The report combines these for a 38% 'active or strategic use' figure, which sounds like progress. In reality, strategic use is still rare.
Adoption rose from 61% to 79% in two years. Strategic use barely moved, from 2% to 4%. The gap between those numbers is what matters.
Even the 79% figure changed during the survey. Early results showed 88% AI usage, but as more small charities responded, the final number dropped to 79%. This isn’t a problem with the data. It shows the report is transparent about its sample. In my view, it also shows that the adoption number is less solid than headlines suggest.
Adoption tells you who has access. It doesn’t tell you if it’s of value.
'Uses AI' is a low bar. If someone opens ChatGPT to draft an email, they count. The number says nothing about whether that use is planned, measured or tied to a goal.
The report’s own progress data backs this up. Eighty-one per cent of charities say they made digital progress this year, a big jump from 60% in 2025, but only 6% describe that progress as “significant,” down from 14% the year before. Around half made only “some progress,” describing the changes as small. More charities are moving. Fewer are moving far.
This isn’t a uniquely British pattern, either. In February 2026, Virtuous and Fundraising.AI published the 2026 Nonprofit AI Adoption Report, a benchmark study of 346 mostly US-based nonprofits. It found 92% of organisations now use AI, but just 7% report a major improvement in what their organisation can actually do. Sixty-five per cent describe their use as “reactive and individual”. That means one person, one prompt, no shared workflow. Only 7% have folded AI into their goals, budgets and KPIs. Virtuous calls this the efficiency plateau: adoption has stalled at faster drafts, quicker emails, same results, because the tools changed but the organisation around them didn’t.
Both reports, in different countries, show the same pattern. Most organisations use AI casually. Only a few use it with purpose.
Why adoption is outpacing strategy
Three things in the Charity Digital Skills Report 2026 explain why adoption is racing ahead of strategy:
The skills to direct AI, not just prompt it, are missing. Fifty-six percent of charities now cite limited skills or technical expertise as their biggest AI barrier. This is up sharply from 43% the year before. As more people start using AI, more people are hitting the edge of what they know how to do with it.
No one is in charge of AI. Forty-four percent of charities are taking no action to progress with AI or manage its risks. This isn’t a training gap you fix with a webinar. It’s a lack of ownership. Without someone deciding what good AI use looks like, staff make their own choices, each in their own way.
Trust in AI is falling. Thirty-five per cent of charities say they don’t trust the AI tools they use, up from 15% last year. The number is even higher for large charities and those serving marginalised groups. Adoption is rising while confidence drops. This happens when people use tools faster than they can check if they work.
Put together, we know more staff are using AI, with less direction, less trust and less skill to make it count. The 79% headline gives us a single reassuring number.
What the strategy gap costs charities
The adoption and strategy gap has a cost, and it’s in the same report. Sixty-three per cent of charities say finances are their biggest barrier to digital progress, and only 17% accessed digital-specific funding from trusts and foundations this year, down from around 30%. Every pound spent on an AI subscription that nobody measures is a pound that isn’t going toward the charities that can point to what AI actually changed.
In the boardroom, if a trustee asks what the AI spend delivered, most charities can only say that people are using it. That’s a weak answer for funders who are tightening budgets.
How the 4% use AI differently
The Virtuous report highlights what separates the 7% (their version of the UK’s 4%) from the rest. These organisations have written governance, shared workflows, and built AI into a few specific, measured processes.
We’ve written about what this looks like in practice: grant pipelines, multi-platform comms, knowledge search and automated impact reporting, all covered in The Agentic Charity. If 79% shows where most charities are, that guide shows where the 4% are.
Three Questions Worth Asking Before You Buy Another AI Tool
Before you buy another tool or licence, ask these three questions:
Where is AI actually used in our organisation, and who made that decision? If the answer is 'wherever staff have picked it up,' you’re in the 79%, not the 4%.
Can we name one thing AI has measurably changed this year, e.g. hours saved, conversion rate, reporting turnaround?
Does anyone own AI as part of their role, with the authority to set how it’s used and report on it? If not, you have adoption without ownership.
To fix AI in your organisation, you need someone senior enough to set direction, connect the dots between fundraising, ops and comms, and hold the whole thing to account.
That’s what a fractional head of digital does. I bring the strategic layer that turns adoption into the kind of measured, embedded use the report’s 4% have already found. It’s also worth reading Why digital work gets harder as charities grow, since the same lack of strategic ownership that stalls AI is usually the same gap stalling digital strategy more broadly. Also read Why generic AI content is quietly killing charity web traffic, which shows what happens when that same ungoverned, individual AI use spills out into public-facing content.
Common questions about the Charity Digital Skills Report 2026
Is 79% AI adoption among UK charities a good sign?
It shows the sector has access and basic use. On its own, it says nothing about whether that use is directed, measured or creating value. The Charity Digital Skills Report 2026’s own figures show only 4% of charities have reached that stage.
What’s the difference between AI adoption and AI value?
Adoption measures how many people or teams have started using a tool. Value measures whether that use is tied to a specific goal, produces a measurable outcome and is repeatable rather than one person’s personal habit. The 2026 report’s “strategic use” figure (4%, page 13) and the Virtuous report’s “major improvement” figure (7%) are both attempts to measure the value side, and both land far below adoption figures.
How can a charity move from adoption to strategic AI use?
Start by naming the one or two processes where AI could change a number leadership already tracks, e.g. grant conversion, reporting turnaround, response times. Rather than letting use spread informally across the team, assign clear ownership, write down basic governance and measure before and after. This is the approach set out in The Agentic Charity.