Where could AI fit in your company? A clear answer in two weeks.
The most common reason UK businesses give for not using AI is that they can't see where it would help. The AI Direction Review finds where AI earns its keep in your business, where it would only add risk, and what to do first.
Led by Dom Roberts, 30 years building software and running a business.
Why start here
"We don't know where it fits" is the biggest barrier.
The barrier isn't usually budget or technology. It's knowing which problems AI is good at in your particular business, and which ones it isn't.
of UK businesses not using AI say they have no identified need for it, and 60% point to limited skills or knowledge.
DSIT, AI Adoption Research, 2026of UK small business owners say they can't see a clear use for AI, and 44% worry about security and privacy.
Simply Business, SME Insights Report 2026of UK SMBs have set no AI budget for 2026, while 31% plan to invest. The gap between them is a plan.
IONOS and YouGov, SMB survey 2026The two weeks
What happens, day by day.
About six hours of your leadership team's time in total, plus short conversations with the people who do the work.
Leadership interviews
Where you want the business to be in two years, what's holding it back, what you've tried with AI, and what worries you about it.
Walk the work
Short sessions with each team to see how work really gets done, plus a look at your systems, data and, for software businesses, your codebase and delivery process.
Score and test
Every opportunity is scored on value, feasibility, risk and readiness. The top one or two get a quick working prototype on your own data, so the plan rests on evidence, not optimism.
Board session and report
A plain-English report and a session with your leadership team: what to do first, what to stop, what it will cost and who owns it.
How opportunities are ranked
Four questions for every idea.
The same scoring is used for every opportunity, so you can see why one idea beat another and re-score new ideas yourself later.
| Measure | The question | What scores well |
|---|---|---|
| Value | What changes if this works? | Hours saved each week, revenue won, errors or risk removed, in pounds |
| Feasibility | Can today's AI do this with our data? | The data exists, is reachable, and a prototype gets close on day one |
| Risk | What happens when it's wrong? | A person reviews the output, or a mistake is cheap and easy to spot |
| Readiness | Will people use it? | A named owner, a team that wants it, and a process it fits into |
What you get
Everything you need to decide, and to start.
- A ranked opportunity mapEvery idea scored, with the reasoning shown
- A 90-day planFirst projects, owners, costs and milestones
- Working prototypesThe top one or two ideas, tested on your data
- An AI usage policy, draftedApproved tools and rules for client and company data
- Build, buy or wait adviceWhich off-the-shelf tools fit, and what is worth building
- A board sessionPlain English, with time for every question
Fit
Who this is for.
A good fit if
- You run an established business, usually 10 to 250 people
- The board or owners are asking for an AI plan
- People are already using AI tools informally, without shared rules
- You want a clear yes, no or not yet before spending more
You may not need it if
- You already know the use case and need it built. See Team productivity or Agentic engineering
- You want a vendor to sell you a platform. I don't resell tools
Questions
The review, answered.
How do I know if my business is ready for AI?
Most established businesses are readier than they think. If you have repeating work, data in systems you can reach, and someone willing to own a change, there is usually at least one place AI pays. The review tells you where, and also when the honest answer is not yet.
Do we need clean data before we start?
No. Part of the review is finding out which of your data is good enough for which ideas. Many useful first projects work with documents, emails and reports exactly as they are.
Is AI worth it for a business our size?
Smaller firms often benefit fastest, because one automation can return a noticeable share of a small team's week. The review puts a value on each idea so you can judge for yourself.
What if the answer is not yet?
Then you'll know why, what would change the answer, and you'll have a usage policy to keep informal AI use safe in the meantime. That's a good result too.
Are we obliged to use you for the work afterwards?
No. The plan is written so your own team or another supplier can deliver it. Any follow-on work is scoped and priced separately.
Related
Before you decide
Talk to Dom
Want a clear answer on AI?
Tell me what's prompting the question. I'll reply within one working day to arrange a short first call.