AI draws attention

Harry at UKcharitycamp

The attention AI brings can be harnessed to improve services and processes

Digishift

Digishift is a series of events by SCVO (the Scottish Council for Voluntary Organisations). For their end of year get-together, they asked:

They were kind enough to get me along to try answering. Here’s what I had to say.

UKCharityCamp

I’m responsible for UKCharityCamp. It’s an unconference where the agenda is set by questions attendees want to talk about. So I have some data.

I dumped the agenda into a word cloud generator and got a list of how often terms show up. The top five words in 2024 in reverse order were “small”, “charities”, “AI”, “lunch”, “digital” (obviously, there were several places you could eat).

This year the running order was “using”, “tools”, “lunch”, “digital”, “AI”. In 2025, AI was more popular than lunch. 

FOMO vs purposeful experimentation

Mor Rubinstein ran a session at UKCharityCamp called – in the immortal words of Rachel Coldicutt – “FOMO is not a strategy”. Rachel blogged about it here and Mor’s session talked it over. 

My hope is that the digital change we’ll see in 2026 is a shift to purposeful experimentation. Not FOMO, but FAFO (feel around and find out).

I’m using the phrase “feel around” because it’s worth spending some time on prep. Don’t just experiment with AI without checking who’s gone before you. Ask around on channels like the “Third Sector Geeks” WhatsApp group. Or check CAST’s library of experiments.

Let’s not waste time – and energy consumption – as a sector repeating the work of others; let’s build on what someone already knows. Only one person needed to prepare X-ray diffraction images of the crystal structure of DNA to reveal it was a double helix. (And her name was Rosalind Franklin.) 

We know specific tasks where AI is useful – it’s great at making rubbish first drafts. You can practise your technique on well-understood tasks, based on what other folk have discovered.    

If you’re doing something genuinely new

If you’re lucky enough to have time and space to genuinely do something new, you need to know the area. Because Generative AI quite literally makes everything up. You need to know if the responses you’re getting are accurate, because it doesn’t. It just knows that they’re statistically reliable. 

Use a structured approach, such as CAST’s AI Experimentation canvas. Then write it up and share, so the rest of us can learn. 

AI and the attention economy

What’s really interesting for me is that AI is clearing up in the attention economy. You need to go back to social media and web2.0 to find the last move of this magnitude. To give you an idea of how long ago that was; it’s before my son was born, and he’s now in his second year at uni. 

You can harness that attention. Think of a niggle – tell folk you want to use AI to ease it, and you’ll get the permission, maybe a day or so to do it, and maybe even the budget to pay someone else to.

Real examples of harnessing that attention

“We’re wasting time reading .pdf files attached to requests for funding. How might we use AI to do that for us?”

Turns out, with a well-structured CRM, and a bit of coding, you can save days a year using AI to do the reading and either grab a bit of data from a file, or tell you it can’t so you can.

“We’re wasting time organising dozens of interviews, for folk from all over the country who can make some times and not others, and are interested in some subjects and not others, who need to be interviewed by an expert in the field, and preferably one who lives nearby so we can minimise our travel budget. How might we use AI for that?”

Turns out, that’s a specific case of the “generalised travelling salesman problem”. A classic computer science problem that can be analysed precisely and the actual, genuine, best answer reliably produced with an algorithm.

Those are real examples from a project we did for the Martingale Foundation. 

Cutting through the hype

Someone rather flatteringly said we’d “cut through the hype and surfaced practical, often surprising ways AI can help teams in the charity sector work more effectively, deliver smarter services and free up capacity.”

Well… use GenAI to get a first draft out fast, but make sure it’s an area you understand, or you won’t know where it’s got the detail wrong.

Use it to drag a specific piece of data out of a structured document, but put the right guardrails in place, so it tells you when it’s not completely sure of its answer. 

And use it to get organisational attention on pain points in your processes. You may not end up using AI, but you’ll likely get the space, time, and maybe even money to ease them.

The best solution to the second question above didn’t use AI, but the word “AI” got us the time to develop it.