The AI opportunity audit

A one-week AI audit of your business. £2,500. A report on day 7.

Under three hours of your team's time. You get a clear picture of how your work actually runs, and what incorporating AI would concretely mean for it — which parts are better done by people, which can be automated, and where the two need to meet.

I have spent the last year building AI systems that run in production for real businesses, so I know what they do well and where they stop being useful. That is what makes the answer specific rather than another version of “you should look at AI”.

One week

intake Monday, report day 7

£2,500

fixed, agreed before we start

Under 3 hours

of your team's time, total

Paid 50% on booking, 50% on delivery. A fixed price, not a day rate — so an audit that turns out harder than expected is my problem, not your invoice.

Why start here

“Incorporate AI” is a phrase, not a plan.

Every business is being told it should be using AI by now. Very few are told what that means for their business specifically: which of the work it would touch, which of it carries on exactly as it does today, and what actually changes on a Tuesday morning. The instruction arrives at the level of the category. The decision has to be made at the level of your operation.

So the audit is a translation. I spend a week understanding how the work in your business really moves, and hand back a clear picture of which parts of it are genuinely better done by people, which can be handed to a machine, and where the two need to meet. It is not a verdict on how you run things. It is a specific answer to a question that is usually only ever asked in the abstract.

The core of it

Which work is which

Not a list of AI opportunities with a disclaimer attached. One picture of how the work divides, with the thinking shown against each line, so you can weigh it yourself rather than take it on trust.

Machine work

Runs without a person, and should

Work that only involves judgement in the sense that somebody has to be awake for it. Retyping between systems, reformatting to a standard, chasing the same status update. Usually the work your most expensive people resent most.

Hybrid

Where most real answers land

The machine drafts and a person signs off — with the handover point specified precisely, because that is the part everyone gets wrong. Vague hybrids collapse into one of the pure answers within a year.

Human work

Automating it costs more than it saves

Judgement, relationships, accountability, and anything where being confidently wrong is worse than being slow. Some of it looks mechanical from outside, which is why it is worth working out properly rather than assuming.

Leave alone

Not now, and possibly not ever

Work that already runs fine, or where the change costs more than the problem, or where the ground is moving too fast to commit. Saying so is part of the job.

The reason I can answer that at any useful resolution is that I have had to decide it in software, repeatedly, and get it wrong first: a value is never guessed at, a document is never sent until a person has signed it off, and the system is never allowed to treat its own output as evidence it was right. Each of those is a line between what a machine should do and what a person should, and each one exists because the alternative broke something.

Setting expectations

What this isn't

Worth saying up front, because the category is full of all four.

  • A strategy deck
  • A maturity score
  • A vendor shortlist
  • A sales document for follow-on work

The deliverable

A written report on day 7, walked through on a call

Five sections. Two of them exist specifically to be useful to you whether or not you ever work with me again.

01

Where the time goes

The recurring work across your business, in your own people's words, and quantified.

02

The allocation map

Every recurring workflow placed as machine, hybrid, human or leave alone, with the thinking shown per line and scored on value, effort and reversibility. This is the section that answers what incorporating AI would actually mean here.

03

The top three

The three moves worth making first, specified concretely enough that anyone could action them, with honest effort and cost estimates you can plan against.

04

Do-it-yourself quick wins

Things you can implement this month without help from me or anyone else. Giving these away is what makes the rest of the document credible.

05

Data and governance note

Including what your staff are already doing with AI unsanctioned, often with client data. Owners rarely know. It usually lands as a genuine finding.

The week

Under three hours of your time, and that is a design constraint

Not an estimate — engagements die of homework. Here is the whole week, and everything it asks of you.

Day 0

A short intake questionnaire — facts only, about fifteen minutes — plus a request for a handful of real working documents.

Owner or office manager

Day 1

A 90-minute call to understand the business as a whole, with a screen-share walkthrough of how the work moves.

Owner

Day 2

A 30-minute call with someone who actually does the day-to-day work. The best findings come from this one.

Someone in operations

Days 3–5

Analysis and drafting, with the occasional short email question as gaps appear.

Nothing needed from you

Day 7

The written report, delivered and walked through on a 45-minute call.

Owner

I ask for documents, not descriptions

A real CV, a client report, an invoice, a typical email thread. Two minutes of your time, and worth more than an hour of answers — people describe a process as it should be, and documents show what it is.

What you provide, in full

  • About fifteen minutes on the intake form
  • A handful of real working documents
  • Ninety minutes of your time
  • Thirty minutes with one person who does the day-to-day
  • A few short email answers mid-week
  • Forty-five minutes for the walkthrough

What I look at

Six areas, the same six every time

Where time actually goes

Recurring tasks by role rather than by org chart. What gets done twice, and what gets retyped from one system into another.

Document and communication flows

What arrives, in what state, what gets produced from it, and who actually reads the result.

System landscape

What you run on, what is connected, where data sits siloed. This determines feasibility more than anything else does.

Bottlenecks and decisions

What waits on one person, and what sits in a queue because nobody owns the next step.

Constraints

Client requirements, regulatory limits, and how comfortable your staff actually are with the technology.

Shadow AI

What people are already doing unsanctioned. This is frequently the finding an owner did not expect and most needed.

How I work

Three rules I hold myself to

Most audits don't end with me building anything

The report is the product. If the right answer is an off-the-shelf tool, a process change, or hiring a person, that is what it will say — and those answers turn up more often than a build does.

I only offer to build what I am confident I can deliver

If something the audit turns up is outside what I know I can do well, I will tell you so and say who or what to look at instead. Taking on work I would be learning on your time is how these projects go wrong.

The audit is complete in itself

It should be worth the fee if you never hire me again. Nothing is held back, there is no phase-two cliffhanger, and the report is yours to act on in-house or with anyone else.

Fit

Who this is for — and who it isn't

A good fit if

  • Profitable, but capacity-constrained — more work than you can handle
  • Expensive people doing manual, repetitive work
  • Growth blocked by admin rather than by demand
  • Document-heavy operations, roughly 10–100 staff
  • No in-house IT team to hand this to

Probably not, if

  • You want someone to validate a decision you have already made
  • You know exactly what tool you need built — book a scoping call instead
  • Nobody can spare a few hours to talk to me

The audit pays off fastest where the pain is a bottleneck rather than a cost line. If you are cutting back rather than straining at capacity, this is probably not where your money should go right now — and I would tell you that on the call rather than after the invoice.

The audit

Get started

Book an audit

Tell me roughly what your business does and where you think the time goes. If I do not think an audit would help you, I will say so instead — that conversation costs you nothing.