Hi there,
Last week we said the way to start with AI is one job, done well. Fair enough, a few of you replied. Which one?
Good question. It’s the difference between AI that quietly saves you an hour a week and AI you try once, get a mediocre answer from and never open again.
So this week, the jobs worth handing to AI first and a simple way to pick yours.
This week in AI
The short version. Markets had a wobble this week over whether all the money pouring into AI will ever pay off. A Gartner study found that firms swapping staff for AI agents often see no return at all. Easy to read that as gloom. It isn’t, not for you. The lesson underneath it is the opposite of the panic. You don’t need a moonshot or a headcount cut. You need one repeatable job done a little better. Which is exactly where we’re headed.
The practical bit
Don’t start with a project. Start with a loop.
Most AI efforts die as projects. Big, vague, ambitious, quietly abandoned three weeks later. The ones that stick are smaller and a lot more boring. You take one task you already do every week and turn it into a loop.
Four steps. You can start the first one today.
- Pick one recurring task. Boring, weekly, low-risk. The status update, the meeting notes, the first draft of the email you send fifty times a month.
- Brief it well, once. Use last week’s habit, role, context, what good looks like. Then save the prompt so you’re not writing it from scratch every time.
- Add a human check. You read it before it goes anywhere. Always. The AI does the first eighty per cent, you do the last twenty and the judgement.
- Count the time. Roughly how long it took before, roughly how long now. Saving an hour a week, you’ve found a keeper. If not, drop it and try a different task.
That’s the whole system. Not a transformation. One task, one loop, one honest measure of whether it’s worth it. Do that with a single job this week and you’re ahead of most businesses spending a fortune trying to do everything at once.
Free resources
What Makes a Good AI Use Case (and What Doesn’t)
Most failed AI efforts didn’t fail because AI doesn’t work. They failed because someone picked the wrong job to start with, too big, too vague, too dependent on perfect data, then decided AI wasn’t for them. This free guide gives you four simple criteria to tell a good first use case from a dead end, before you spend any time or money on it.
New here? You’ll set up a free community login on the way in.
What’s on
The sessions below ran in June 2026. For what is coming up, see upcoming events.
Nine days to go. In one morning you walk out with your AI strategy and policy written by you. That’s the Build Along, on 9 July. There’s a free afternoon session too, so come for one or both.
Build Along with AI
9 July 2026 · White City, London · 11:00 to 13:00
Walk in unsure about AI. Walk out with an AI strategy and a one-page policy you wrote yourself, in one morning. A working session with us in the room, not a talk.
Free virtual waitlist
The AI Foundry Live
9 July 2026 · White City, London · 13:30 to 16:30
Our quarterly free live event. Two short talks from guest speakers, then time to compare notes with other UK business owners over an afternoon coffee.
Speaking this time: Mahdi Shariff from Humble AI and Hannah Morris from Let’s Trailblaze.
Free, uncapped
When you’re ready to go further than one job, this is the room to do it in, with us there to help. Bring the person who really owns this in your business.
The giants are still proving ROI. You just need one less thing on your plate.
The biggest companies in the world spent this week being asked a very simple question. Is any of this actually paying off? You can answer that one by Friday. Pick a job, do it well, count the time you saved. That’s a return, a real one. Most of the market is still hunting for theirs.
Next week, the one habit that stops AI quietly embarrassing you. The human check that keeps it honest.
Hit reply and tell me the one weekly job you’re going to hand over first. I read every reply.
Know someone drowning in the same repetitive task every week? Forward this on.
Understand first. Then act.
Richard