What is an AI audit, and what do you get for £950?

Why bother with an audit?

Because the expensive mistake isn’t building something badly. It’s building the wrong thing well.

Without an audit, automation projects tend to start from whatever tool someone has just read about. With one, they start from where your team’s time actually goes. The second approach is cheaper, because you only build what pays for itself.

What I look at

The audit is about your work, not about AI. I’m looking for:

Where the hours go. Which repetitive jobs take up the most time each week.
What’s worth automating. Which of those jobs pass the tests in which jobs are worth automating: frequent, consistent, rule-based, costly when wrong.
What isn’t. Jobs that need judgement, happen rarely, or would be better fixed than automated.
Your existing tools. What you already pay for, and whether it can already do some of this.
Your data. What personal or sensitive information each job touches, and therefore what kind of AI, if any, is appropriate. (Where your data goes explains the options.)

What you get

A written report you keep, whatever you decide. It sets out:

the jobs worth automating, in order of how much time or money they’d save
the jobs that aren’t worth automating, and why
for each recommended build, a fixed price and a timeline
which AI approach each job needs: a commercial service, a model on servers you control, or no AI at all

That last point matters. Plenty of automation needs no AI. If a job can be done with plain rules, it should be: it’s cheaper, faster and easier to check.

What it costs

£950, fixed.

If you go ahead with a build, the £950 comes off the build price. So if the audit leads to work, it costs you nothing extra.

For comparison, one UK consultant’s guide puts a proper audit for a business under ten staff at £750 to £2,500. What AI automation costs covers builds and running costs too.

What if it finds nothing worth automating?

Then you keep the report and the fee stands.

That’s a genuine outcome, not a failure. Knowing that your processes are already lean, or that the obvious candidate wouldn’t pay back, saves you from spending thousands on something that doesn’t earn its keep.

What an audit isn’t

It isn’t a sales pitch for a build. The report is yours. You can take it to another developer, or build it in-house.
It isn’t a strategy deck. It’s a list of specific jobs, with specific prices.
It isn’t about replacing people. It’s about finding the copying and checking that takes up your team’s day, so they can spend that time on work that needs them.

Is it worth it for a very small business?

If you’re a sole trader, possibly not. The most useful thing you can do is the week-long time log described in which jobs are worth automating. If something on that list takes hours every week, then an audit starts to make sense.

For a team of five or more, where several people spend part of each day on admin, the audit fee is usually small next to what that admin costs.