Most businesses that have had a disappointing AI project hired the right skills for the wrong problem. They needed someone to tell them what to build and why. They hired someone who was very good at building. The technology worked. The commercial outcome did not materialise. That is not a technology failure. It is a hiring failure that was avoidable.
The Core Difference
An AI developer builds what they are told to build.
An AI consultant decides what is worth building and why.
That distinction sounds simple. In practice it determines whether an AI investment delivers commercial value or produces an impressive prototype that sits unused.
A developer optimises for technical correctness: does the system work as designed? An AI consultant optimises for commercial outcome: does the system deliver the business result it was built to deliver? The two are not the same thing, and the gap between them is where most AI projects fail.
What an AI Developer Does Well
An AI developer is the right hire when the commercial problem is already clearly defined, the solution is already designed, and what you need is someone to build it correctly.
They work from a specification. They select the right models and tools for the job. They write the code, build the integrations, and ensure the technical system functions as intended. They can advise on technical trade-offs: which model, which infrastructure, which approach to a specific integration problem.
What they are not trained to do is ask whether the thing they are building is the right thing to be building. That is not a criticism. It is not their brief.
What an AI Consultant Does Well
An AI consultant is the right hire when the commercial problem is not yet clearly defined, when previous AI projects have not delivered expected results, or when a significant investment is being considered and the business wants an independent commercial view before committing.
A consultant maps the existing workflows and identifies where AI can deliver measurable value. They build the commercial case. They design the solution around the outcome, not the technology. They own the adoption plan that determines whether the team actually changes how it works. They define the ROI metrics before anything is built and track them afterwards.
A commercially-led AI consultant can usually spot within the first conversation whether an AI project is being scoped from the right direction or the wrong one. Most are scoped from the wrong one.
How to Tell Which One You Need
There are four signals that tell you which gap you have.
Signal one: You have a clear brief with defined outcomes. If you can describe the specific workflow being automated, the baseline metrics, the target outcome, and the success criteria, you probably have enough commercial clarity to brief a developer. If you cannot, you need a consultant first.
Signal two: A previous AI project did not deliver what was expected. This is almost always a commercial gap, not a technical one. The technology worked. The wrong thing was built, or the right thing was built for the wrong problem, or the team did not adopt it. A developer will not diagnose or fix any of those problems. A consultant will.
Signal three: You are trying to justify AI investment to a board or leadership team. A developer cannot build the commercial case. They do not have the commercial framing or the business case methodology. If you need to get a project approved internally, you need a consultant to build the case before a developer builds the system.
Signal four: You know you want AI but are not sure where to start. This is the most common situation and the clearest signal that you need a consultant. Starting with a developer when you have not yet defined the problem leads to something being built, often something technically impressive, that does not solve the commercial problem you actually have.
Why Hiring a Developer When You Need a Consultant Is Expensive
The cost of the mistake is not just the developer's fees. It is the full cost of the project: the build time, the integration work, the team time invested in the rollout, and the opportunity cost of the months spent on the wrong thing.
A typical mid-size AI project that does not deliver commercial value costs between £20,000 and £60,000 in direct spend, plus the internal time of the people involved. More damaging in many organisations is the scepticism that a failed project generates. Leadership teams that have seen one AI project fail to deliver are harder to convince on the next one, even when the next one is properly scoped.
The investment in commercial clarity before the build begins is small relative to the cost of getting the build wrong.
When You Need Both
The best-run AI projects have both. A commercially-led consultant owns the brief, the business case, the adoption plan, and the commercial outcome. A developer or technical team builds what the consultant has designed. The two work together, in that sequence.
The failure mode is when those roles are collapsed into one person who is stronger on the technical side, or when the sequence is reversed and the build happens before the commercial design.
If you are evaluating which to hire first, hire the consultant first. The consultant will tell you what to build. The developer will build it. The commercial thinking has to come before the technical execution, not after.
A Practical Test
Before your next AI project begins, answer these five questions:
What is the specific commercial problem we are solving? Not "we want to use AI." A named process with a quantified cost.
What does measurable success look like in twelve months? Not outputs or features. Commercial outcomes.
Who owns the adoption plan? Not the build. The behaviour change.
What is our baseline? The before state against which results will be measured.
What is the cost of doing nothing for another year?
If you can answer all five with confidence, you are ready to brief a developer. If you cannot, you need a consultant first.
Frequently Asked Questions
What is the difference between an AI consultant and an AI developer? An AI consultant decides what to build and why, and owns the commercial outcome. An AI developer builds what they are told to build, and owns the technical correctness of the system. Both are necessary for a successful AI project. The consultant comes first.
Can an AI developer also be an AI consultant? Some technically experienced people have developed strong commercial thinking and can operate effectively in both roles. They are rare. The more common pattern is a developer who presents as a consultant but defaults to technical framing when it matters. The test is whether they ask about your commercial baseline and success metrics before they recommend any technology.
Is an AI consultant the same as a fractional CTO? No. A fractional CTO typically focuses on technical strategy, engineering team management, and architecture decisions. A commercial AI consultant focuses on business outcomes: where AI can create measurable value, how to build the case for investment, and how to ensure the team adopts what is built. The roles can overlap but they are distinct.
When is the right time to hire an AI consultant? Before you have decided what to build. The earlier a commercial consultant is involved, the less likely the project is to go in the wrong direction. The most expensive mistake is hiring a consultant after a developer has already built something that is not working commercially.
How do I evaluate an AI consultant before hiring them? Ask for the commercial outcome of a previous project, not the technical description of what was built. Ask how they define success before a project begins. Ask what their adoption plan looks like. Ask what happens if the AI underperforms. A good consultant has specific answers to all of these. A developer presenting as a consultant usually does not.
What does a commercially-led AI consultant charge? A focused commercial audit or assessment starts at around £3,500. A full project engagement covering audit, solution design, commercial case, and oversight of the build typically ranges from £15,000 to £60,000 depending on scope and team size.
Jessica Thomas is a Commercial AI Consultant who works with B2B companies and agencies to define the commercial problem before the technical solution. If you are trying to work out which type of support you actually need, [book a call](https://calendly.com/jess-jessicathomas/30min).
Ready to stop talking about AI?
If you're done with frameworks and workshops and want someone who can actually deliver — let's talk.