AI implementation: one real task, working in about four weeks.
For small and mid-sized companies that have tried ChatGPT and seen nothing change. We pick one task your team does every week, build it on your own documents and rules, and hand it over with an owner and a person checking the output.
AI implementation, done well, starts with one task rather than a strategy. I spend a few days finding the task worth doing first, then about four weeks getting AI to do it in daily work. In week 1 we pick it with the people who do it and record how long it takes today. In weeks 2–3 I set AI up on your own documents and rules and test it with your team. In week 4 your people use it on real work, with a named owner and a person checking every output. Every step is a fixed fee, quoted in writing.
In PwC's 29th Annual Global CEO Survey (January 2026, 4,454 CEOs), 56% said AI had brought neither higher revenue nor lower costs in the past 12 months.
Does this sound like your company?
Everyone's trying ChatGPT. The work, the costs and the results look the same as last year.
IT says six months. You need it this quarter.
You paid for a strategy. Nobody built anything.
Your best people spend their week on quotes, reports, emails and document checks.
What's usually missing is someone whose job is getting the tools to work. That's what I do, one task at a time.
Step one: find it (the AI opportunity review)
Best if you're not sure where to start.
The review takes a few days. I hold short interviews with the people who do the work and look at where the hours and errors actually go. You get a ranked list of opportunities, with rough cost and time for each, and the first project scoped. I put it all in a written plan and walk you through it. The plan is yours to keep, whoever does the work.
The review also includes the 90-day test: which parts of your business a well-funded competitor using AI could plausibly rebuild in 90 days, and what you could now build yourself.
Fixed fee, agreed before we start.
How we workOpportunity review or readiness assessment?
They answer different questions. A readiness assessment scores how prepared your company is for AI. An opportunity review finds the tasks worth doing, ranks them with rough cost and time, and scopes the first project, so you know exactly what to build next. If you already know you want to use AI and need to decide where, the review is the more useful of the two.
Optional quick win
If you use ChatGPT Business or Claude, or want to, the review can include setting it up for one team, with shared templates, a one-page acceptable-use note and a short handover. It's quoted in writing as an add-on.
Step two: the first project (about four weeks)
A first project takes one task from choosing it to your team using it every day. At the end, it's part of daily work.
Week 1, find it
I spend half a day with the people who do the work. We pick one task worth fixing and check what it depends on. We record how it's done today, including checking and corrections, so we have something to compare against. We also agree what the AI may see and do, and where a person decides. You'll have a named owner on your side from day one, and we agree their time in the proposal.
Weeks 2–3, build it
I set it up on your own documents, templates and rules. Then I test it with your team on real examples, including the awkward ones. The missing detail is flagged, never invented.
Try the enquiry example (made-up)Week 4, hand it over
Your people use it on real work, with a named owner, written instructions and a person checking the output. We compare time per task before and after, and you decide whether to do the next one.
Exact scope and timing are agreed in writing once we've checked what the task depends on. Fixed scope and fixed fee. No hourly billing.
If week one shows the task isn't a good fit, we agree the next step together, in writing.
What you have after four weeks
- A working tool on your documents
- A named owner and written instructions
- A person checking the output
- Time per task before and after, including checking and corrections
You end up with something your team uses every day, rather than a strategy deck or a pilot that lives on one person's laptop.
You own the code, workflows and data built for your business. Account ownership, licences and support are set out in the proposal.
Got a task in mind? Bring it to a free 15-minute call.
Good first tasks
A good first task happens often, follows clear rules, uses your own documents and is quick to check.
Typical first projects (illustrative)
Sales
Call notesa first-draft proposal on your templateyour salesperson checks and sends
Finance
Invoices matched to purchase ordersmismatches flaggedpaid only after a person checks
Operations
The month's figuresa monthly report draft, with gaps and unusual values flaggedthe owner checks it before it goes out
Customer service
An enquiry in Germana reply drafted from your own answerschecked before it's sent
People and HR
A staff question about a policya draft answer with the source pageyour HR team checks it
Contracts
Outgoing contracts checked against your current termsdifferences flaggedyour team decides
Each one keeps a person in charge of the result. Yours will depend on where the hours actually go, which is what week 1 finds out.
Not a good first project: anything customer-facing without review, or decisions that need a licensed professional.
How I make sure it's used, and right
- Your people test it on real work before it goes live.
- The people who'll use it get handover training.
- A person approves anything that leaves the business.
- Automatic checks run first, then I look at the result myself.
- We agree the measure up front: time per task, before and after.
I work the same way on bigger builds. For a yacht brokerage (client, live), I built the website and broker portal. Before each release, the 240+ listing pages, the images and the brochures are checked automatically, and then I check them myself. That was a larger build, not a four-week first project.
Listing pages: live sitemap count, 24 September 2026.
How a yacht brokerage keeps 240+ pages rightWhen to go further
Most companies start here. You only move up when this step is working.
- If several systems need to stay in step: workflow automation and integration
- If your rules are ones no tool can follow: custom AI and software development
- If you need someone to own AI over time: a fractional Head of AI
- Find it
- Quick win
- One workflow, end to end
- Connected systems
- Custom AI tool or agent
- Platform or product build
- Run and improve
Questions
How do I implement AI in my small business?
Start with one task your team repeats every week, not a company-wide plan. Measure how it's done today, set AI up on your own documents and rules, test it with the people who do the work, and hand it over with an owner and a person checking. Then compare before and after.
Do we need a technical team?
No. I set up and build what's needed, directing AI to do the technical work. Your people bring what matters most: how the work is really done. You'll need a named owner on your side, and we agree their time in the proposal.
Do we need to choose the AI tools first?
No. Start with the work you want to improve. I'll recommend tools once we know the task, and we use what you already have where it does the job. If you already pay for ChatGPT Business or Claude, we start there.
How long does it take?
The opportunity review takes a few days. A first project takes about four weeks. Exact scope and timing are agreed in writing once we've checked what the task depends on. Larger work is split into phases, each with its own timing agreed in writing.
How will we know it worked?
We record time per task before we start, including checking and corrections, and compare it after week 4. You decide what's next based on that comparison. The measure is agreed with you up front, so it's clear from the start what we're comparing.
What does AI implementation cost?
The review and the first project each have a fixed fee, quoted in writing before we start. The proposal also states software costs, who owns the accounts and what support is included after handover. There's no hourly billing, and nothing starts until you've agreed the quote.
What about our data?
We agree up front what data the work needs, what the AI may see and do, and where a person must decide. I work with OpenAI, Anthropic (Claude) and xAI (Grok), all US providers, chosen for the task. Ask me how your data would be handled for your specific task.
What if AI isn't the answer?
Then I'll say so, before we agree to build anything. Sometimes the fix is a better process or a simpler tool. If week one shows the task isn't a good fit, we agree the next step together, in writing, before anything else is built.
Is an opportunity review the same as an AI readiness assessment?
Not quite. A readiness assessment scores how prepared you are. An opportunity review finds the tasks worth doing, ranks them with rough cost and time, and scopes the first project. You come away knowing what to build first and roughly what it will take.
Who owns what you build?
You own the code, workflows and data built for your business. Account ownership, licences and support are set out in the proposal, so it's clear before we start which accounts and third-party software are yours and who looks after what.
“I don't know very much about technology, but I can certainly see the excellent result, and the care you've put into helping [the business] grow.”
Tell me one task your team hates doing.
In a free 15-minute call, we'll work out whether AI can take it on and what a sensible first step would be. If the answer is no, I'll say so.
Free · 15 minutes · you'll talk to me, not a sales team.
