The diagnostic
Two weeks. Fixed price.
Honest answer.
Where AI would help your business, where it wouldn't, and what the numbers actually look like. One page, written for a board, delivered at the end of week two.
20-min call first. We'll tell you if it's worth doing.
What it is
A short, paid engagement designed to answer one question.
Should you be doing AI, where, and what's it worth? We interview your team, look at your processes and data, and write you a one-page report. Not a vendor pitch in disguise. Not the start of a fourteen-month transformation. Just a clear answer.
About one in five diagnostics ends with us recommending you don't automate — or that you fix something else first. If that's where it lands, we refund the second week. We'd rather lose the build than sell you something that won't work.
Week one
Discovery.
Two of us on the ground or on calls. James in the leadership sessions, an engineer alongside on systems and data.
-
Leadership session
Half-day with whoever signs off the budget and whoever owns the affected processes. We map the operational pain, the data you have, and the constraints we need to work inside.
-
Operator interviews
Three to six conversations with the people doing the actual work. Not just heads-of. We want to know what gets done twice, what relies on one person, what gets fudged.
-
Data and systems walk-through
We look at the inputs an AI system would need. Where does the data live? Is it usable? What would have to change before any pilot could run?
Week two
Analysis and write-up.
We disappear, we think, we write. You'll see us again at the end of the week for the walk-through.
-
Mapping to what works
We test each opportunity against what current AI tools can actually do in production — not what a demo suggests they could. Mature patterns: document workflows, internal retrieval, classification with judgment. Less mature: anything that needs novel reasoning or unstructured negotiation.
-
Numbers
For each shortlisted opportunity: rough cost to build, rough cost to run, expected payback. Conservative on saving estimates, explicit about assumptions, so you can stress-test them in the boardroom.
-
Write-up
A one-page report. Prioritised opportunity list, yes/no recommendation on the top three, and the rationale you can take to a board without further translation.
What you walk away with
One page. Board-ready.
Acted on, not filed.
01
Prioritised opportunity list
Three to seven candidates ranked by impact and feasibility, with the assumptions behind each ranking written down.
02
Yes / no on the top three
For each of the top three candidates: do this, don't do this, or do this once X is true. With a written rationale you can defend.
03
Numbers and timeline
Rough cost to build, cost to run, expected payback, and a sensible delivery timeline. Conservative on savings. Explicit on what we don't know yet.
The refund mechanic
If, after week one, the honest answer is "don't do it" — we tell you, write up what we found, and refund the second week.
This isn't a get-out-of-jail clause we hope never to use. About one in five diagnostics end this way. The point of the engagement is the answer, not the sale. If the answer is no, we'd rather you have a short report and your money back than a strategy deck dressed up as progress.
A good fit if
-
You have three or four processes that quietly cost a salary each
-
You're being asked by the board to 'do something about AI' and want an independent view
-
You've been pitched and the proposals contradict each other
-
You've tried AI before and it didn't stick
Probably not, if
-
You already know exactly what to build and just want a quote — go straight to Build & Deploy
-
You're under three people and you're not sure you've got the process volume to justify it
-
You want a strategy deck for board theatre — that's not what we produce
What it's not
A hundred-slide deck. A pitch in disguise. The beginning of something larger you didn't ask for.
The diagnostic is a self-contained piece of work. You can take the report, thank us, and never speak to us again. That's a perfectly reasonable outcome and we've had it happen.
If we do recommend a build and you'd like us to do it, we'll quote it as a fixed-price engagement starting from the report. No surprise scope. No "phase 2" appearing in month four. The diagnostic is what you paid for; the build is a separate decision.
Not ready for a diagnostic yet?
The half-day AI Briefing is the cheaper way to find your feet — what current tools can and can't do, what to ask vendors, where to look in your business.
Twenty minutes. No deck.
A short call first. We'll talk through what you're looking at, give you an honest read on whether a diagnostic is worth doing, and roughly what it would cost.