Chief AI officer or head of AI: which role, and when

The titles get used loosely, so start with what each person answers for. A chief AI officer answers to the chief executive or the board for AI across the whole organisation. A head of AI answers for making it work: the team that builds, the platforms it runs on and the systems in production. One sets direction and carries the risk, and the other ships. In a smaller organisation one person does both, which is fine as long as the mandate says so.

The clearest written description of the executive job I know comes from the US government. In April 2025 the Office of Management and Budget (OMB) told every federal agency to retain or designate a Chief AI Officer, positioned highly enough to engage regularly with the agency's deputy secretary. The memo's list of duties reads like a decent job description for any large organisation: promote AI adoption through a governance and oversight process, act as senior adviser on AI to the head of the agency, keep an inventory of AI use cases, put processes in place for high-impact uses, advise on making the workforce AI-ready, advise on AI investments and support efforts to track AI spending.

Look at what is missing from that list. Nothing in it builds anything. Building is the head of AI's job, and confusing the two is how an organisation ends up with a strategist expected to ship, or an engineering leader expected to run the board's conversation about risk.

The third column is the one people skip. The OMB memo itself lets an agency designate an existing chief information, data or technology officer as its Chief AI Officer, provided that person has significant expertise in AI. If a government that requires the role will accept an existing executive in it, a private company does not need a new title to take AI seriously.

Public bodies are in a different position, because some of this is written down for them. In the US, OMB memo M-25-21 (3 April 2025) requires every federal agency to have a Chief AI Officer. At the largest agencies, those covered by the Chief Financial Officers Act, the post must sit at Senior Executive Service level or an equivalent senior grade, and elsewhere at or above grade 14 of the General Schedule. Those largest agencies must also convene an AI governance board chaired at deputy secretary level or equivalent, with the Chief AI Officer as vice-chair.

UK guidance takes a different route. The government's Digital and Data Profession Capability Framework defines a chief data officer, a chief technology officer and a chief digital and information officer, and has no chief AI officer role. The AI Playbook for the UK Government does not call for one either. It asks for an AI governance board, or AI expertise on an existing board, to set principles and review and authorise uses of AI, and for a senior responsible owner who is accountable for the use of AI in each project. For a UK public body, the practical answer is usually to give the AI mandate to an existing chief, then make the board and the project owners real.

Private companies have no rule either way. The US National Institute of Standards and Technology (NIST) AI Risk Management Framework, a voluntary framework, states the principle plainly: executive leadership takes responsibility for decisions about risks associated with AI system development and deployment. Whoever holds the title, that responsibility has to sit with someone senior enough to carry it.

I would not create a chief AI officer title to show the board that AI is being taken seriously. A title without decision rights, a budget and the authority to stop things is a spokesperson. The role earns its seat when several of these are true.

A head of AI is the right first hire when the strategy is broadly agreed and the gap is delivery: use cases picked, budget approved and nothing in production. That person needs a team or a supplier to lead, and an executive sponsor who already sits at the leadership table.

You need neither yet when AI use is a handful of off-the-shelf assistants run through their vendors' admin controls, or one function running one pilot. Name an executive owner, keep a register of the tools in use and put a short policy in place. Our guide to measuring AI ROI shows how to tell whether that early spending is paying back, which is the evidence a bigger role will later be justified on.

A permanent senior hire takes time to find and to land, and the decisions do not wait for it. There are four ways to cover the gap, and they cover different halves of the job, so be clear which half you are buying.

Whichever you choose, plan the exit on the first day. Interim cover with no end date becomes the operating model by default, and the permanent hire then inherits decisions nobody wrote down.

The job advert is the last thing to write. Settle the mandate first, because candidates will ask about it and the good ones will walk away from a vague answer. For either role, decide who they report to, which decisions they may take alone, which budget they control, what they can stop and how the first year will be judged.

Then test for judgement. These are the questions I would ask any candidate for either role, and I would pay more attention to how specific the answers are than to the vocabulary in them. The brief below puts the mandate and the questions on one page you can send before the first interview.

This page quotes no salaries. I could not find a primary source that publishes private-sector pay for these roles, and pay tracks the scope of the mandate more than the title. Benchmark against your own leadership team for comparable scope, or ask a search firm for current data.

Two of our offers map onto this decision. If nobody owns the technical side of the AI programme, our Fractional AI Implementation Team does the head of AI's delivery work while you settle the permanent structure. We own the target architecture and the vendor and model decisions, build the platform and the first production workflows alongside your engineers, and keep a dated decision log. It is retained, typically two to three days a week under one statement of work, and it is designed to end with your own team running what we built. For a large international marketing agency, we hold the lead architect role on its AI platform while the agency's own leadership sets the direction, which is the split this page argues for.

If the executive question comes first (where AI should pay back, and what it will take), the AI Opportunity & Feasibility Sprint answers it in a fixed scope, typically two to four weeks, and ends in a roadmap and an investment case that a new chief AI officer or head of AI can pick up. If you already have a scoped job, whether a small fixed-scope statement of work or a larger build, we can resource it on contract from the collective of associates who work with 1AYM, held to the same standard. If you would like to talk through who should own AI in your organisation, book the call below.