Comparing AI consultants by their rate is the worst way to choose one. Compare them on five things instead: what will be different on a Tuesday three months from now, who owns the system when the engagement ends, what they will refuse to do, what they have built for a business shaped like yours, and how you will know it worked in ninety days.
Gartner expects over 40 percent of agentic AI projects to be canceled by the end of 2027, because of escalating costs, unclear business value or inadequate risk controls. The prediction ran in a press release dated June 25, 2025. None of those three reasons is a model problem. They are scoping and judgment problems, which is the work a strategy consultant is supposed to do. The rate tells you what the engagement costs, but not whether anything will change.
Five questions to ask before you hire
Here are the things to ask, in this order.
Start with what will be different on a Tuesday three months from now. Not launch day. A normal working day. Installing a tool changes your software. Changing how work runs changes who does what, in what order, and what stops landing on your desk. An answer worth trusting names one real task and says how it will go.
Next, ask who owns the system when the engagement ends. A good answer is concrete. The files that tell the system how to work are the business’s, you can open and read them, and you can leave with a copy. If the answer is a subscription that ends your access when you stop paying, keep looking.
Then ask what they will refuse to do. A consultant who takes every job has not told you what they are good at. Ask what they would tell you not to do. The honest answer is specific, a little uncomfortable, and sometimes costs them the work.
Ask what they have built for a business shaped like yours. Skip the logos. Ask them to describe a build in your kind of work: the judgment involved, the confidentiality, the standards nobody ever wrote down. They should not name the client. If they can describe it well without the name, they understood it.
Last, ask how you will know in ninety days. Pick one measure before the work starts. Hours spent on a single task. How long a client waits for a reply. How many drafts you rewrite. One number, written down, agreed on.
A consultant who builds this ends up holding how you work. That means how you price, what you do when client work goes badly, and the standards you never wrote down. You spent years learning them. When I hear that the business owns the system, I want that to cover all of it, in a form the owner can read and carry out the door.
The rate tells you what the engagement costs, but not whether anything will change.
The comparison you are making is usually not consultant against consultant. It is consultant against a hire. A consultant usually costs less and leaves sooner. That is a good trade when the work is scoped to something that stays after they go. It is a bad one when you are paying for a relationship instead of a result.
Ask the five questions before the second call. If the answers are vague, that is your answer.