Boutique or Large Consultancy? An Honest Comparison

The premium is not margin. It buys specific things, and you may not need them.

A large consultancy is not charging four times more for the same work. It is selling a different product, and almost none of the extra money goes into the code. It goes into professional indemnity cover, a balance sheet that will still exist in five years, a bench that can replace your lead engineer in a fortnight, a name your board will accept without argument, and the capacity to put twenty people in four countries at once. Those are real goods. The honest question in the boutique versus large AI consultancy decision is whether your project needs any of them, because if it does not you are paying for insurance against risks your project does not carry, and if it does, no small firm substitutes at any price.

What the two kinds of AI consulting firms actually cost

AI consulting firms cluster into two commercial models, and the numbers below are published European market ranges rather than anybody’s price list. They move with country and sector. Three figures matter more than the headline day rate: what a day costs, what the smallest engagement they will accept costs, and what it costs simply to get a scope written.

• Day rate. A large firm bills a blended team at roughly €1,200 to €2,500 a day, with partner time well above that. A specialist firm bills senior engineers at roughly €600 to €1,100. The gap is smaller than most buyers expect, which is why the day rate is the least useful of the three numbers.
• Minimum engagement. This is where the real difference sits. A large consultancy’s commercial model rarely works below about €150,000, because the cost of selling, staffing and governing an engagement is close to fixed. A specialist will take €3,000 to €10,000 of work and still make it worthwhile. If your project is a single workflow, you are not choosing between two suppliers. Only one of them can accept the job.
• Discovery. A large firm’s scoping phase commonly runs €25,000 to €75,000 and produces a considerable document. A specialist’s runs €2,000 to €8,000 and produces a shorter one. Both are legitimate. The difference is that at the first price the discovery is a decision in itself, and at the second it is a cheap way to find out you should not build.

Which brackets those map onto by project type is in our note on what a custom AI build costs, and the shape of the contract you sign in either case is covered in fixed fee, day rate or outcome.

Five things enterprise AI consulting buys that a small team cannot

We are a specialist firm, so read this section as the version we would rather not write. All five are real, and each one has cost small suppliers work they were technically capable of doing.

Risk transfer with a balance sheet behind it. Professional indemnity at a level that means something, and an entity large enough to still be there when a claim is made. On a system that touches money, safety or regulated decisions, your legal team is buying the counterparty as much as the software, and they are right to.

Continuity. If the lead consultant leaves, a large firm replaces them next month from a bench. When a four-person supplier loses the person who wrote your system, you have a problem no clause solves. That is the single strongest argument against the small end and it deserves a plain answer rather than a deflection.

Governance cover. Some decisions need a name that a board, an auditor or a regulator recognises. This is not a technical requirement and it is entirely real. A director who has to defend a choice in a committee is buying defensibility, and telling them it is irrational does not make it go away.

Procurement eligibility. Public tenders and large corporate frameworks routinely require a supplier revenue floor, a certification, or three years of audited accounts. A specialist firm fails the filter before anyone reads the proposal. Check the qualification criteria before you shortlist anybody, because it decides the shortlist for you.

Simultaneous scale. Twenty people, four countries, one timeline, with change management in each market. A small firm would subcontract that, and you would pay a margin for the coordination without getting the coordination.

What a boutique AI consultancy trades, and what you accept in return

The advantages are the mirror image, and they are mostly about who is in the room. At a small firm the person who scoped the work is the person writing it, so nothing is lost handing over. A question asked on Tuesday morning is answered by Tuesday afternoon rather than routed through an engagement manager. Decisions that would need a steering committee happen in a call. On a twelve-week build that difference is worth more than the day rate, because the schedule is set by the pace of decisions and not by the pace of typing.

Memórias do Jamor is a fair sample of what that buys. A public fan wall for Sport Clube União Torreense, open to anonymous uploads, machine-triaged before anything publishes, with a cost failsafe that degrades the automation rather than the service as usage climbs toward its ceiling. The interesting decision there is not technical, it is the ordering: survive the launch spike first, bring the running cost down second. That call was made by the people building it, in a conversation, in an afternoon. Read the full build and the rest of our case studies and you will see the same pattern, which is the honest reason we can move quickly.

What you accept in exchange is key-person risk, a smaller safety net, and the fact that we cannot be everywhere at once. A serious boutique AI consultancy reduces the first of those rather than denying it: documented code you own, a repository in your organisation rather than theirs, an architecture note a new engineer can read in an hour, and the arrangement written down for what happens if the firm becomes unavailable. Ask for all four in writing. A supplier who treats the question as an insult has answered it.

Senior sells, and the ratio of junior hours decides delivery

One dynamic runs through this market and appears in no proposal. The people who sell the work are frequently not the people who do it, and the gap widens with firm size. It is structural rather than dishonest: a firm of four hundred cannot put its most experienced person on every engagement, and the commercial model depends on a specific ratio of junior to senior hours. It is still your risk, and you can price it with one question, asked identically of both types of supplier. Which named individuals will do this work, what else are they committed to this quarter, and will you name the key person in the contract?

A large firm will often decline the last part for legitimate staffing reasons, and a good one says so plainly rather than agreeing and then substituting. A small firm that declines it has a different problem, because there is nobody else who could be doing the work. Either way, the answer tells you what the engagement will feel like in month three, and it is worth more than any slide in either pitch.

Deciding by project size, company size and risk profile

Below roughly €150,000 of project value, the question does not exist. A large AI consultancy cannot serve that engagement profitably and will either decline it or staff it thinly. Choose among specialists and spend your energy comparing them rather than comparing categories.

Above that, ask what fails if the supplier disappears. If the answer is a project delay, a specialist is fine and cheaper. If the answer is a regulatory exposure, a public incident or a stalled multi-country programme, you are buying continuity and indemnity, which is precisely what enterprise AI consulting charges the premium for.

Then check whether procurement or the board has already decided. A revenue floor in the tender or a director who needs a recognised name has settled the matter before the technical comparison starts. Find that out in week one, not week six. The wider version of this sort, across five supplier categories rather than two, is in the five types of AI partner, which is the page to read first if offshore studios or an in-house hire are also on your list.

Where we sit: we are a specialist firm and we take work between roughly €3,000 and €50,000, senior people on the build, code and documentation handed over at the end. We are the wrong call for a multi-country rollout, for a board that needs governance cover, for a procurement process with a revenue floor, and for anything an existing product already solves. In a meaningful share of first conversations the recommendation is smaller than what the client asked for, and occasionally it is that they should be talking to a large firm instead. That answer costs us the project and it is still the right one to give. What that first phase looks like in practice is set out in what an implementation actually involves.