What AI consulting really costs

“How much does it cost?” is the first question most leaders ask us, and they’re right to. AI consulting prices vary widely, and the way a project is priced often matters more than the headline number.

This article explains what drives the cost, the common ways engagements are structured, and what we charge.

What you are actually paying for

Almost every AI engagement is made of the same four kinds of work. The price depends on how much of each your situation needs.

1. Diagnosis. Understanding your business, your data and your systems well enough to choose the right use case. This is mostly senior time: interviews, analysis and a written recommendation.
2. Build and integration. Making the use case work inside your real systems. Connecting to a CRM or ERP, handling permissions and dealing with messy edge cases usually takes more effort than the AI itself.
3. Data work. Cleaning and connecting the data the use case depends on. This can be small or large, and it’s the part most often underestimated.
4. Adoption. Training, documentation and changing how a team works. Skipping this is the most common reason a working tool goes unused.

When two quotes differ a lot, it is usually because one includes integration, data work and adoption, and the other assumes you will handle them.

How engagements are usually priced

Hourly or day rates. You pay for time. It’s flexible, but the total is open-ended, and the risk of overruns sits with you.

Retainers. A fixed monthly fee for a set amount of access or capacity. Useful for ongoing support, less useful for getting a specific result.

Fixed scope and price. You agree what will be delivered and what it will cost, in writing, before work starts. The risk of overruns sits with the consultant, so the scoping has to be careful.

Outcome-linked fees. Part of the fee depends on a measured result. This sounds attractive, but it only works if the baseline and the measurement are agreed upfront, and they rarely are.

How we price

We work on fixed scope and price, agreed in writing before we start.

• The Growth & Operations Diagnostic is a fixed fee. It takes two weeks and ends with your baseline, the top changes ranked by value and effort, and a 90-day plan. If you go ahead, the fee is credited to your project.
• A sprint (AI Pilot-to-Production, Growth or Automation) is priced on scope. It takes four to eight weeks and ends with one change live, measured against the baseline.

The sprint price depends mainly on the integration and data work the change needs, which is why the diagnostic usually comes first. You know the number before any work begins.

Five questions to ask any AI consultant

1. Is the price fixed, and what exactly is included? Ask specifically about integration, data clean-up and training.
2. Who will do the work? Ask for the names of the people on your engagement, not just the people in the pitch.
3. How will the result be measured? If there is no baseline, there is no way to know whether it worked.
4. What happens if it doesn’t work? A good answer includes the point at which the consultant will tell you to stop.
5. Who owns what gets built? You should own the configuration, code, prompts and documentation.

The cheapest project is the one you stop early

The most expensive AI projects aren’t the ones with the highest fees. They’re the ones that run for months without a clear measure of success. A short diagnostic that tells you not to build something can save far more than it costs.

If you want a quick view of where your company stands before talking to anyone, take the AI Readiness Index. It’s free and takes about six minutes.