Who should own AI in a UK business

Barclays has given its first Chief AI Officer role to its group chief of staff, Anita Tanna. Her background is equity sales at Barclays, Citi and Goldman Sachs. It is a business appointment, made from inside the firm, and it puts AI ownership next to the chief executive’s office.

The same month brought two other UK answers. Mewburn Ellis promoted its long-serving IT director into a new CTO role with AI in the remit. JTC merged the technology job and the AI job into a single Chief Technology and AI Officer. Three firms. Three different seats for the same question: who should actually own AI?

Clario’s view is that any of the three can work. The choice that tends to fail is creating the title before the business has decided what that person is accountable for. A Chief AI Officer who cannot stop a project, a CTO who is handed AI on top of an already full job, and a combined title with no agreed split all run into the same problem. The board has a name. The company still has no owner.

A business operator

Some companies want the owner to come from the commercial side. They already have a technology leader. What they lack is someone who can sit with the chief executive and the business units and decide which AI work continues. Anita Tanna’s move is that kind of choice. The technology team still has to be in the room, because a commercial owner who cannot judge the systems will sign off work they do not understand.

This shape fits when AI is already in several parts of the business and the gap is authority, not a shortage of engineers. The brief has to say which decisions move to the new role, and which stay with the CTO. If both people believe they own the plan, the firm has hired a dispute.

The technology leader

Other companies keep AI with the person who already runs technology. Mewburn Ellis widened an IT director role into a CTO post and put AI inside it. That can be the right call when the CTO has the time, the board’s confidence, and a team that can do the work. It is a poor call when the CTO is already the only person who can keep the systems running, and AI becomes a slide in a board pack.

The test is the week, not the title. If the technology leader can name the AI projects that should stop, explain the cost to the board, and still run the existing estate, one role may be enough. If those tasks would land on top of an unchanged job, the firm needs a second person, usually a Head of AI reporting to the CTO.

One combined title

JTC chose a single Chief Technology and AI Officer. A combined title can suit a firm that wants one executive, one team and one budget. It asks a great deal of one person. The brief should say what is in scope this year, and what can wait.

What to settle before the search

Before any of these titles goes to market, the brief needs four sentences. Who the person reports to, and who can overrule them. What they can stop or fund in the first year. Whether the data has an owner. What the salary and the location are.

Clario writes that down with the chief executive, and with the CTO when that is a different person. You review it before anyone is approached. A title chosen for a board paper, with the accountability left until later, is the version that stalls.

The firms named above are public appointments. They show the range of serious answers. The useful question is who will be allowed to decide, and what they will be judged on in eighteen months. These roles usually pay £100,000 or more, and more when the person sits with the executive team. The Chief AI Officer and CTO pages set out the split, and the guide to hiring a first Chief AI Officer is written for boards that are still choosing.