How to choose an enterprise AI consultancy in the UK

These four checks are answerable from production systems, not from a demonstration. The figures below are already published on our engagement files, each with its provenance. 1AYM works with clients across the UK, the US and the Gulf.

Search for an AI consultancy in the UK and the results mix four kinds of firm that share a label and almost nothing else. Which one you need depends on the problem, and most bad engagements start by buying the wrong kind, not the wrong firm.

If you are still deciding between an agency and a consultancy, our explainer on what an AI agency is and how it differs from a consultancy covers that question first, including when an agency is the better buy.

Two more kinds of supplier turn up on real shortlists. One is the vendor's own team, where your contract includes deployment help. The other is the firm that sells itself as an AI development company or an AI agent builder, a label common in US searches for the same purchase. Each kind has one question that separates the good ones from the rest, and it is rarely the question the pitch was rehearsed for.

Four of the six offer coding-agent rollouts: the vendor's own team, specialist consultancies, large systems integrators and contractors. If the job is rolling out Codex or Claude Code across engineering, ask each candidate whether it builds the operating layer around the tool (shared instructions and skills, permissions, review gates and evaluation in CI) or only switches on seats and runs a training day. Measure adoption by merged changes that pass review. A count of licences measures spend.

No UK body publishes a genuine benchmark of consultancy fees. The Management Consultancies Association publishes market size rather than prices, Source Global Research publishes pricing sentiment with the detail behind a paywall, and the Crown Commercial Service publishes its consultancy framework's grade structure and the fact of contractual caps without the figures. The fee guides that circulate online cite nobody, which is why their numbers disagree with each other.

The only public, firm-attributable prices in the UK are the per-day fee schedules suppliers must publish to sell through the government's G-Cloud framework. G-Cloud 14 is in force until October 2026, with a successor announced in August 2026, and most of its cards were priced in April and May 2024. Read them as a floor rather than a quote: ONS figures put producer price inflation for professional, scientific and technical services at 3.5% in the year to the first quarter of 2026.

Suppliers price against SFIA, a seven-level seniority scale, which is what makes the cards comparable. The figures exclude VAT and assume an onshore person-day with professional indemnity cover included. At level 5, the senior specialist grade, published prices run from £800 to £2,380 a day across the eleven supplier fee cards read for this page on 1 September 2026, with a median of £1,500. The large firms sit at the top of that level: Accenture, Deloitte and KPMG publish £1,580 to £2,380 at level 5 and £1,880 to £2,855 at principal or partner level. Suppliers outside the large firms publish £800 to £1,400 for the same level, with Fusion AIX at the floor of that band. Whole engagements carry fixed prices too. On its G-Cloud 14 card Gartner prices whole engagements rather than days, publishing £70,200 for AI use case elicitation and prioritisation, £142,400 for an AI capability maturity assessment and £203,500 for an AI strategy and roadmap.

What surprises buyers in the catalogue is that the large firms do not simply price higher. Across whole listings their published bands are wider at both ends, typically holding lower floors as well as higher ceilings, and the floor is where the misreading happens.

None of which makes a published figure a quote. The useful question is not the price of a day but what a fixed scope delivers and who carries the risk when the work runs long. 1AYM sells three shapes: fixed-scope statements of work, a retained implementation engagement, and engineers embedded in a client's own programme. None of them is priced on this page, and the embedded engagement states its commercial terms on its own page. Four things in the same public record are worth more to a buyer than the bands themselves.

The single most predictive question in selection is whether the people who design the system stay accountable for building it. Where strategy and engineering are separate departments, engineering quietly rewrites the strategy, and systems drift from the deck that sold them. Where the people who design the work answer for the build, there is no handover to hide behind.

Ask to meet the people who would deliver as well as the people who sell, and ask what they personally shipped in the last quarter, with named systems.

Demonstrations are cheap and production is expensive, so production is the evidence that matters. Ask every candidate for one real workflow they took from ambition to a system people use daily, and how long it took.

Then read the numbers with their provenance attached. A figure published by the client, a figure the consultancy measured, and a forecast are three different strengths of evidence, and a firm that blends them into one claim is telling you how it will report your programme too.

Every firm will say the word governance. The separating question is where it lives. If the answer is a policy document and a committee, the controls depend on people remembering them. If the answer is identity and permissions the system enforces, deterministic checks that gate what an agent may do, evaluation that runs on every change and human approval on consequential actions, the controls hold when nobody is watching.

The pattern to look for is the model proposing and something deterministic deciding: an agent drafts, and validators, reconciliations and human gates decide what proceeds. A consultancy that cannot describe its gating pattern in that level of detail has not built one.

Most AI consultancies hold some vendor relationship, and the honest ones state it plainly, because it shapes advice. A reseller attached to one vendor will recommend that vendor. A model-selective firm chooses by workload and production evidence, and can show you systems running on more than one stack.

Partner statuses are worth reading precisely. 1AYM, for example, is an OpenAI Select Partner: a company status that describes a relationship with the vendor, stated so a buyer can weigh it. Treat any partner badge the same way, from any firm: as a disclosure to interrogate, not a shortcut past the four checks above.

The end state of a good engagement is that the client owns the code, the architecture, the operating model and the capability, and could continue without the consultancy. Ask directly: what do we own on the last day, who can run it, and what does it cost to keep running.

Firms whose commercial model depends on you being unable to leave will resist that question. That resistance is the answer.

Ownership is easiest to check early. Our guide to the first 30 days of an AI implementation sets out what each week should produce once you have hired, including code and prompts in your own accounts from the first commit.

A pitch can pass the four checks on paper and still carry warning signs. Any one of these should slow you down, and two should end the conversation. Each has a check that takes minutes, and a firm that resists the check has already answered it.

These are buying checks, not legal advice: have your own counsel read the contract before you sign.

In the US the same purchase is often searched as an AI development company, or an AI agent development company when the thing to be built is an agent. The label promises code rather than advice, which is a good start, but the four checks still decide it: who writes the code, what runs in production, where the controls live and what you keep.

Hiring one well follows a short sequence. Write down the problem, the systems it touches and how you will measure success before you speak to anyone. Shortlist three firms from references and work you can inspect, and treat a directory's order as a starting list at most. Put the questions in the table above to each, then buy a small fixed scope before a large one: one workflow, an evaluation set built from your own cases, and a handover. Our template sets out the clauses to put in an AI statement of work, including ownership of code, prompts and test sets.

An agent development company needs harder questions, because an agent acts rather than answers. Ask what the agent may touch and how that is enforced, how every action is logged, where a person approves before anything consequential happens, how the agent is tested against your cases on every change, and what it will cost to run each month. Ask what has to happen to your data and infrastructure first. True project costs are usually clear only after a discovery stage, because sorting out data and infrastructure often comes before the agent build, and our guide sets out the controls and cost drivers of an agent build.

US federal buyers already have this discipline in writing. OMB memorandum M-25-22 tells agencies to test proposed AI solutions before award where practicable, to write in protections against vendor lock-in such as knowledge transfer, data and model portability and rights to code and models produced under the contract, and to bar vendors from training commercial models on non-public agency data without consent. A private company can ask for the same terms in its own contract.

If the problem is a well-served product category, buy the product. If you have strong engineering leadership and a clear architecture, hire or borrow engineers instead of buying design you already have. If nobody senior owns the outcome internally, fix that first, because no external firm can substitute for an absent owner.

If you are not sure which of these describes you, score your organisation with our AI readiness assessment first: 13 questions on ownership, data, systems and controls, with the result shown on the page.

The choice between your own team, contractors and a firm usually turns on what you already have.

A consultancy earns its fee where strategy, architecture and engineering have to move together and the organisation wants to keep the result: taking the first governed workflow into production, standing up the platform and controls around it, and transferring the capability to run it.

The questions that come before and after choosing a supplier, such as which AI plan to buy and how to govern it once it is in use, each have a page of their own in our guides to buying AI services.

If you are choosing a supplier for a first production system, 1AYM is one of the firms you could put through these checks, and we would rather you did. Where the question is still what to build, the AI Opportunity & Feasibility Sprint settles it in a fixed scope of typically two to four weeks. Where it is an agent or a workflow, Enterprise AI Platform & Agentic Workflows designs and builds it with the controls this page asks about.

For teams rolling out coding agents such as Codex and Claude Code, AI Engineering Transformation builds the operating layer around them. 1AYM is an OpenAI Select Partner, and we build on both vendors' tools. If you already have a scoped job, we can resource it on contract from the collective of associates who work with us, held to the same standard.

We take small fixed-scope statements of work as well as larger builds. 1AYM will tailor the project to your budget and timeframe, and your account manager will make sure expectations are set from the first conversation. The next step is a 30-minute call, or an email with your shortlist or draft scope.