You don't need a head of AI. You need someone who owns the decision.

In spring 2025, Deloitte surveyed 550 business and technology leaders on AI decisions. Where the CFO had no decision-making authority, only 18% achieved above-average profitability. Where the CFO had full decision-making authority, 42% did. That is an association, not proof, and Deloitte does not say how many firms sat in each group. What went with better results was an existing director holding the decision. For most companies of 50-350 people, the answer is not to hire a head of AI. It is to make one existing director answerable for what AI is for, with a budget, a number and the authority to cross departments.

The title is spreading

In a year, the chief AI officer has gone from unusual to normal. IBM's 2026 CEO study, a survey of 2,000 chief executives and senior leaders, found that "76% of surveyed organizations have a Chief AI Officer (CAIO) in 2026, up from just 26% in 2025". That is what the companies say about themselves. In the UK, ITJobsWatch counted 144 permanent vacancies for a Head of AI in the six months to 11 October 2026, against 29 in the same period last year.

The title is associated with better returns. IBM's 2025 study of chief AI officers found that "Organizations with CAIOs see 10% greater ROI on AI spend". The same study found that "57% of these CAIOs were appointed from the organization's internal talent pool". In most of those companies, the chief AI officer was someone already on the payroll, handed the decision. That is ownership under another name.

A title is the easiest part of an AI strategy to copy.

Where AI lands when nobody owns it

Without a named owner, the decision tends to default to whoever runs the technology. Info-Tech's survey of 551 senior leaders, published in July, found that "CIOs and CTOs remain the most common owners of AI initiatives". Tools get bought and pilots run, and nobody answers for what the business should do differently. Info-Tech also found that organisations with a dedicated, governed AI strategy were "three times as likely to report measurable AI impact", at 60% against 20% of those with no active strategy. IT can run the tools; it cannot decide what sales, finance and operations should stop doing.

Most of this evidence comes from large firms. The UK figures show the mid-size picture, and it is earlier still. The government's AI adoption research, a survey of 3,500 businesses, found 23% of mid-sized firms, those with 50 to 249 staff, currently using AI. In its follow-up interviews, "very few businesses said they had any internal guidelines or specific policies around using AI responsibly or ethically". NatWest's study of 1,400 small and mid-market firms, published on 25 September, found that 44% use AI. Of the firms using it, only 6% had reached the most advanced stage.

AI is arriving in mid-sized companies faster than anyone is deciding what it is for. Buying AI seats before anyone has named the work is often what that looks like: a decision nobody owns, made one licence at a time.

What the owner needs, and who already has it

Owning the decision means holding the budget, so that spending on AI is a choice and not a request to someone else. It means carrying one number that everyone can check. And it means authority across departments, because the work AI changes rarely sits inside one. That describes a director, not a new hire who reports into one.

At this size, the candidates are already on the leadership team. The COO, where the work AI would change is mostly operations. The CFO, where the case rests on cost and margin. Sometimes the MD, when nobody else has reach across the whole business.

A head of AI without a budget is an adviser with a better title.

The objection comes from Deloitte itself: "No single role can drive strategic KPI alignment, technical capability growth, and profitability on its own." Nobody is asking one director to do the work alone. The owner answers for it: decides what AI is for, sets the number, and is the person the board asks.

IBM found the same shape: organisations where chief AI officers drive a centralised or hub-and-spoke operating model are associated with up to 36% higher ROI on AI initiatives. Decisions sit in one place, and the work still happens in the departments. Without that, MIT Sloan Management Review warns: "Left unchecked, most organizations will expand AI projects and offerings in an uncoordinated manner."

The case for the hire

A dedicated hire makes sense when AI changes what you sell or how your market works, not when it mainly makes current work faster. Info-Tech's study found that organisations with dedicated chief AI officers "report the highest rate of department-wide adoption", and the highest rate of measurable impact. That is a seat with enough AI work to fill it, which at 50-350 people is rare. Writing for the Harvard Law School Forum on Corporate Governance, Fawad Bajwa of Russell Reynolds Associates sets the test as whether a company is "looking to make a market shift", or whether AI is "changing the face of your products and services". He allows that the role can be an anchor hire that draws in AI talent. "But they're also not a silver bullet." MIT Sloan's authors agree: "The right answer depends on the strategic importance and maturity of AI in your company." For most mid-sized UK firms, on the figures above, AI is still a productivity programme, and that needs an owner, not a new seat.

Name one director from your leadership team. Give them the budget, one number and the authority to cross departments. Revisit a dedicated hire only if AI starts changing what you sell.

The owner is not on the market. They are already in the room.