What does an AI consultancy cost in the UK?

Most UK AI consultancies won’t put a number on their website, so here is the answer with our own figures on the table. Our AI Audit and Transformation is a fixed fee from £10,000, delivered within four weeks, with a costed 12 month roadmap at the end. The Embedded AI Lead, the retainer many clients keep on afterwards, runs from £5,000 a month. Published UK market ranges put a single-workflow agent at roughly £10,000 to £30,000, and multi-agent work spanning several systems at £40,000 to £150,000 (OpenKit AI development cost guide). Regulated environments add ten to twenty percent for the security and audit work. OpenKit never quotes from a rate card: we scope each build against a defined outcome first.

The rest of this guide is where each figure comes from and what moves it, because two honest quotes for the same brief can sit far apart once you see what each firm priced in.

What an AI audit costs.

Every OpenKit engagement starts with the audit, and its fee is fixed in writing before anything else happens. What lifts the figure above £10,000 is scope: how many workflows and systems you want in range, and the interviews it takes to map them properly. The output is a written report and a prioritised 12 month roadmap with a cost against each recommendation, which is what makes every later figure in this guide checkable, since each one traces back to it. We’ve written up what an audit looks like from the inside if you want the mechanics rather than the money.

Across the UK market the same work sells as discovery or as a readiness assessment, and it’s usually quoted only after a call, because the work changes shape with every client. A fixed fee settles who carries the risk of a wrong guess about scope. We do, since the fee can’t move once it’s agreed.

What an AI retainer costs.

After the audit, many clients keep us on to deliver what it found. OpenKit becomes your business’s embedded AI team: the people who ran your audit stay on to deliver your 12 month roadmap. That engagement is the Embedded AI Lead, from £5,000 a month, scoped against what the audit surfaced and reviewed month by month, so the commitment renews monthly rather than locking in a year.

The nearest thing sold elsewhere is fractional AI leadership, where UK firms selling the service advertise between £2,000 and £15,000 a month depending on days per week. The overlap in price hides a gap in what the money buys. A fractional executive sells you a person’s days, so the direction is theirs but the build still needs your team or another supplier. We set the routes against each other in our chief AI officer guide.

What a bespoke build costs.

There’s deliberately no OpenKit figure here. A build is costed in the roadmap the audit produces, from what the audit found in your data and systems, so the number is for your build and arrives with the reasoning attached. The market ranges at the top of this page are the way to sanity-check any quote you’re already holding, and what pushes a project toward either end of them is covered in our development cost guide: data readiness first, then how many systems the AI has to touch, then the governance the work carries.

Some roadmap lines are not builds at all. Where an off-the-shelf tool does the job for less, the roadmap says so, and the cost against that line is a subscription. We can say that because the audit does not need the answer to be a build. What we build when a build is warranted is on the AI engineering page.

Day rates versus fixed fees.

Most UK consultancy work is sold one of two ways, a day rate for a team or a fixed fee for a defined outcome, and the day rate is the less informative number even though it looks easier to compare. The total is the rate multiplied by a day count nobody has fixed yet. A low rate across a long, badly scoped engagement costs more than a higher rate across a tight one.

We publish starting fees and no day rates at all, because a rate without a day count says nothing about the bill. If you’re comparing day-rate proposals, ask for the work scoped in days against a defined outcome, then judge the rate against the total it produces.

What moves the price up or down.

For the audit the lever is range, meaning the workflows and systems in scope and the interview count that follows from them. For the retainer it’s the mix inside each month: how much is build work against the roadmap and how much is steering. For builds the drivers are larger and better documented. Data readiness comes first, integration comes second, and governance is the one most quotes understate, because access controls, audit trails and security documentation take real engineering time in a regulated environment.

Prices move down too. A scoping call that removes workflows is doing its job, and the cheapest line in most roadmaps is the one that says buy, or the one that says leave the process alone.

How to budget the first year.

You do not have to commit to the year up front. A free AI readiness check takes minutes and tells you which of the drivers above will weigh on your figure. The audit follows at its fixed fee, the roadmap it produces puts a cost on everything after it, and the retainer renews month by month, so at no point in the first year are you signed up for spend you haven’t seen justified. The fees and what each covers are on the pricing page, and the exact figure always arrives in writing before you sign anything.