AI consultant or fractional CTO: which one does the problem need?
One is hired for a technology and the other for a business. How to tell which you are short of, and what it costs to hire the wrong one.
Both titles turn up in the same search, and both people will take the call. They’re hired to answer different questions, and hiring one to answer the other’s question is where the money goes.
Two briefs, two starting points
An AI consultant is hired for a technology. The brief starts from the tool: we should be using AI, tell us where. That’s a fair brief, and it’s the right one when the business already knows which of its problems is the expensive one.
A fractional CTO is hired for a business. The brief starts from the company. The technology is holding us back, or we can’t judge what we’re being told, and somebody senior needs to own it. AI is one thing that person might reach for. So is a rewrite, a hire, a different supplier, or doing nothing at all for six months.
Ask an AI consultant where AI fits and you’ll get a list of places AI fits, because that’s the job they were hired for.
The same request, on two desks
Say the request is this. We want AI to draft the first reply on support tickets.
An AI consultant takes the sentence as the work, and that’s the correct reading of it. They’ll look at a sample of your tickets, choose an approach, build something that drafts a reply, tell you what it costs per thousand, and show you how it behaves when it can’t answer. Done properly that’s a useful piece of software for a reasonable fee.
I’d read the same sentence as a symptom. Why are the replies slow? How many of the tickets are the same three questions, and what would a decent help page do to the volume? What does a wrong reply cost you? Has the support manager asked twice for a better search on the knowledge base and been told there’s no budget? Then, if the answer still holds up, build the drafting.
What each one leaves behind
An AI consultant leaves you a working thing and a document that explains how it works. That’s what you bought, and for a business that knows what to do with both it’s good value.
A fractional CTO leaves decisions, and people who know why those decisions were made. It’s slower to see and harder to write an invoice against, and it’s still there in two years.
Neither is the better buy; they answer different questions, and you’re the one who knows which of those questions is costing you money.
The question that settles it
Ask yourself what you’d do if the answer came back as “no AI at all”.
If that result would be worth the fee on its own, you’ve got a business problem and you want the person who owns business problems. If it would feel like a wasted engagement, you’ve already decided to build with AI and what you’re short of is a good builder.
The first kind is more common. A lot of the AI projects that reach me don’t need a model at all.
What the wrong hire costs
The condition that goes missing is ownership. A business with nobody holding the technical direction hires an AI consultant, gets a working pilot, and then finds there’s nobody to run it, nobody to say how it fits with the systems around it, and nobody to decide what happens when the model changes underneath it. The specialist has gone, correctly, because the work they were hired for is finished.
The opposite mistake is quieter, and it runs for longer before anybody notices: a fractional CTO who can’t build hands the AI work to somebody else, so the decisions get made in one room and the code gets written in another, and you’re paying two people for one conversation.
When an AI consultant is the right hire
This section is about the consultant and nobody else. There’s no matching list for the other direction, because a business that can write these three lines has already done a lot of what a fractional CTO gets hired to do.
You can describe the workflow you want changed, in the words the people who do it every day would use, and as a sequence. A ticket arrives, somebody reads it, decides which of eleven queues it goes in, pastes in one of four standard replies, and passes the odd one to a manager. A specialist can start from that on the Monday. Without it, the first weeks of the fee go on finding it out, and finding it out is a different job from building.
You’ve worked out what a wrong answer would cost. That number decides how much of the fee goes on checking the output, and a specialist who has to ask you for it halfway through will price the checking as a change.
Somebody inside the business owns the technical direction. The specialist’s work has to land in an architecture that someone is looking after, and on the last day of the engagement there has to be a person to hand it to. Without that you’ve bought a pilot with a leaving date.
With those three in place a specialist is faster than I am, and you shouldn’t pay for breadth you won’t use.
Where I sit, and what I’d do first
I sell both of these, so weigh that. The case studies show what the combination looks like in production, and the about page has the business behind them.
I start from the business either way: what is this work supposed to change commercially, and how would we know it had? A company that already has a strong engineering lead in place might well do better with the specialist.
If you can’t yet say which workflow should change, don’t hire anybody to build. Find out what’s costing you first.