The right AI strategy consultant for a company of a few dozen to a few hundred people sells you a written plan with named owners, not a platform or a team of builders. You can tell within one conversation: ask who owns the plan after they leave, and listen for a name from your own company. This article is for the owner, CEO or operating partner who is deciding whether to hire outside help for the AI plan, and how to choose. What the engagement should look like day to day, and what a consultant should never do, is in what does an AI consultant do.

I was COO and then President of a public company that closed more than 40 acquisitions, and I ran the integration on most of them. Now I write fixed-fee AI plans for CEOs of private-equity-backed and owner-led mid-market companies through 013 Labs. I am a lawyer by training. I sell one of the options described here, so weigh the comparison with that in mind.

What you are actually buying

You are buying a decision on paper. Technology is the smaller part of it.

Picture a distribution company with a few hundred people. The board has asked, for the second time, what the AI plan is. Meanwhile people in the office already paste customer quotes into a public chat tool because it saves them an hour. Nobody decided that, and nobody has written down what must never be built.

That company does not need software yet. It needs four answers in writing: what is already happening, who is accountable for it, what is allowed, and what is never allowed. An AI strategy consultant is worth hiring if that is what they hand you. If they hand you a tool, a demo or a long deck with no owner, you have bought something else.

The three kinds of firms that answer the phone

Search for an AI consultant for business and three kinds of firms come back. Each fits a different company.

The systems integrator who wants to build. This firm's business is software projects, so the proposal is usually a build. That is right when you know what you need built. It is the wrong first call when the question is whether to build at all, because the adviser also earns money from the answer "yes."

The big-firm practice priced for enterprises. These teams bring depth and a strong name, and they are built for companies with large staffs and a budget for months of study. For a company of a few hundred people, the engagement often arrives sized for a bigger organization, with a final report your CEO has to translate into action alone.

The operator who writes the plan. This is a person or small firm that has run a business and sells the plan itself: who decides, who owns, what is off limits, and what to do first. It fits an owner-led or private-equity-backed company that needs a defensible answer for a board, a lender or a buyer without buying a build. The limit is capacity: an operator writes the plan, and you still need people to carry it out.

If you know what to build, hire the builder. If you run a large enterprise, a big-firm practice may fit. If the question is "what is our plan, and who owns it?", start with the operator.

Seven questions to ask before you sign

Ask these in the first meeting. Each has an answer you want.

Who owns the plan after you leave?

Why it matters: A plan owned by the consultant stops moving when the consultant does. The answer you want: A named person inside your company, with a few hours a week of real time and the authority to say no. If the answer is "your steering committee" or "our team will stay on," keep asking.

What will you tell us not to build?

Why it matters: An adviser who has never said "do not build that" is selling, not advising. The answer you want: Specific examples, starting with anything that sends money, signs a contract or makes a promise to a customer without a named person's approval.

What do you need from our people, and how many hours?

Why it matters: A mid-sized company has no spare staff. A plan that quietly needs many people for weeks will stall. The answer you want: A short list: the owner, the finance or operations lead, and a few hours from the people who already use AI well.

How do you charge, and what does the first deliverable look like?

Why it matters: Pricing shapes behavior. Open-ended hourly work rewards a longer engagement. A fixed fee for a defined deliverable rewards finishing. The answer you want: A fixed fee, a stated end date, and a description of the first piece of paper you will receive, in terms a board member who was not in the room could read.

What happens with our data?

Why it matters: Customer records, pricing and employee information are where the legal and reputational exposure sits, and a plan rarely requires system access on day one. The answer you want: A plain rule: what they will see, where it is kept, who else can read it, and what is deleted at the end.

What have you run yourself, not just advised?

Why it matters: Advice from someone who has never carried the result is easy to give and hard to use. The answer you want: A direct account of what they operated, with no names of clients or results you cannot check. If the answer is a list of tools they know, that is a different qualification.

How will the board hear it?

Why it matters: The plan is not finished when the consultant is satisfied. It is finished when your board, lender or buyer accepts it. The answer you want: A written brief in plain English, a clear account of who presents it, and a way to update it each quarter instead of replacing it. The agenda and the ten questions are on one page: AI board briefing template.

Red flags in the first meeting

What a good first engagement looks like

A good first engagement is small, written and easy to stop. A short call confirms the question your board is asking and who should be in the room. Then comes one working session of about ninety minutes with the leadership team. There is no technology demonstration and no pitch for a build.

Within a week you receive a written brief: where AI is already in use in the company without anyone deciding it, the two people on your staff who already work well with AI, with a written scope for each, and the short list of things never to let anyone build. That is the shape of the AI Board Briefing. Then you decide. You may stop there. If you want the full plan, a four-to-six-week strategy sprint turns the brief into a roadmap with owners, measures and a defined first build, and the sprint ends in a board decision.

When you do not need a consultant at all

You can skip outside help in one situation. The company already has the one person colleagues ask when a spreadsheet misbehaves, that person has real time and the owner's backing, and a two-page plan already exists with the owner, the limits and the first step written down. If both are true, you have what a consultant would deliver. The free guide to spotting your AI champion helps you check the first half.

You may also want ongoing leadership rather than a one-time plan. That is a different role, covered in what a fractional chief AI officer does.

Frequently asked questions

What does an AI strategy consultant do?

An AI strategy consultant helps the leadership of a company decide where AI belongs in the business, who is accountable for it, and what must never be built. The deliverable is a written plan with named owners and measures, not software.

How much does an AI consultant cost?

It depends on how the fee is set. A fixed fee for a defined deliverable tells you the cost before you start and rewards finishing. Time-and-materials billing is open-ended: the more the work expands, the more you pay. For a first engagement, ask for a fixed fee in writing along with the end date.

Do we need a chief AI officer or a consultant?

A consultant answers a question once: what is our plan? A chief AI officer, whether full time or fractional, answers it every quarter and sits with leadership while the plan is carried out. Most mid-sized companies should begin with the plan, because it shows what the role would have to do.

Is AI consulting for a small business different?

The questions are the same. The difference is capacity: fewer spare hours and less room for a failed build, so the first step should be smaller and the never-build list stricter. If a proposal needs more of your people than you can spare, it is sized for someone else.

How long should it take to get a plan?

A written first view can take about a week after one working session. A full strategy, with a roadmap, owners and a defined first build, takes four to six weeks. Be wary of any answer that is open-ended.

Not sure where you stand? Take the 3-minute AI readiness scorecard.