Barclays' New AI Chief Shows the Chief of Staff Is Becoming the Chief AI Officer

Barclays has a chief AI officer for the first time, and she has never run a technology organization. Anita Tanna announced on LinkedIn this week that she is taking the role after nearly three years as group chief of staff to CEO C.S. Venkatakrishnan. Before that, she ran equity sales and program trading execution at Barclays and spent 12 years at Citi, finishing there as UK head of equity sales.

We think Barclays made the right call, and more companies should copy it. Once a bank has built its AI platform, the chief AI officer's job is mostly getting people to change how they work. That job should go to someone the division heads already answer to.

Barclays already had senior technologists in charge of AI. In September 2025 it made Antoinette O'Neill group CIO with AI strategy inside her remit. It has a managing director who carries the titles CIO and Head of Artificial Intelligence. By the end of 2025, 89% of its applications ran on cloud platforms, about 70% of core enterprise data had been standardized, and more than 250 AI tools and models were in use.

Most of the infrastructure is in place. In her post, Tanna says she will focus on "accelerating adoption across the group" and on the areas where AI can return the most value. She doesn't mention models, vendors or platforms. For this job, that is the right emphasis.

Builders and Operators

The chief AI officer title has spread fast. In IBM's annual CEO study, 76% of organizations surveyed reported having one in 2026, up from 26% in 2025. Banks have filled the seat in two ways.

The builders came up through analytics and engineering. Lloyds hired Sameer Gupta, chief analytics officer at DBS. UBS hired Daniele Magazzeni, chief analytics officer at JPMorgan. Bank of Ireland hired Prag Sharma, who ran Citi's Global AI Centre of Excellence. Commonwealth Bank hired Ranil Boteju, who led data and analytics at Lloyds.

The operators came from inside the business. HSBC promoted David Rice, COO of its corporate and institutional bank. Barclays has now promoted Tanna straight out of the CEO's office.

HSBC is the closest match. When it named Rice in March, it also expanded CTO Mario Shamtani's remit to cover core platform modernization and a central AI platform. That left Shamtani responsible for the technology and Rice for getting staff to use it. Barclays appears to have divided the work between O'Neill and Tanna in a similar way, though it picked its AI chief from the CEO's office.

Hiring a builder makes sense when the platform still has to be built. Barclays says nearly all of its applications already run in the cloud and hundreds of AI tools are live. Its problem is usage, and it hired for that.

Why a Chief of Staff

IBM's study also helps explain the choice. 83% of CEOs said AI success depends more on people's adoption than on the technology, and they estimated only 25% of their workforce uses AI regularly. Closing that gap means getting division heads to change how their teams work, moving budget, and settling disputes over who owns what.

Much of that is already a group chief of staff's job. Tanna spent nearly three years reporting directly to Venkatakrishnan and working across every division, and she wrote that she plans to bring that experience into the new job. For an adoption push at a bank this size, that may count for more than a technical background.

Barclays also had a gap in ownership. A tracker of bank AI programs noted this month that Barclays named no single chief AI officer, with group co-COOs Craig Bright and Anne Marie Darling fronting the major rollouts. And its group chief data officer since 2020, Luke Ashton, left for the Office for National Statistics earlier this year.

The Cost Program Behind the Title

Barclays is targeting around £2 billion in savings between 2026 and 2028 and has promised more than £15 billion to shareholders over the same window. At the half-year results, Venkatakrishnan said the bank would take further cost actions later in 2026.

Staff will likely tie a chief AI officer focused on adoption, named in the last quarter of that year, to the cost program, whatever the announcement says. HSBC's appointment drew the same connection. It came amid reports that as many as 20,000 roles could be at risk as the bank cuts back its middle and back offices.

Barclays has said it is pacing itself. Darling told American Banker the bank is not using its most expensive model "like a Ferrari, to go around the block."

That restraint is sensible, but it doesn't answer the question employees will ask first. If Tanna's job includes taking cost out, Barclays should say so now. Staff will find out either way, and people are slow to adopt tools they believe are being used to replace them.

Who Controls the Budget

The risk with the operator model is a job with influence but no budget. If the CIO controls the platforms, vendor contracts and engineering spend, a chief AI officer without any of those mostly coordinates. Barclays hasn't said whether Tanna will have her own budget or team, or whom she reports to.

Our view: a chief of staff can make this work, but only with her own budget and a seat on the group executive committee. Without those, the title becomes a coordination role, and Barclays is back where it started, with AI spread across several executives and no one clearly in charge.

What Leaders Should Take From It

If your AI tools are deployed and usage is flat, the fix is probably organizational, and the right person to run it may already work in the CEO's office.

Settle the split between your AI chief and your CIO before announcing the role: who owns the model budget, the vendor contracts, the list of use cases and the adoption targets. HSBC announced both remits on the same day.

Judge the role by weekly usage and cost savings. Barclays counts more than 250 AI tools and models, a figure that says little about how many people use them or what they have saved.

Expect employees to assume an adoption push means job cuts, and tell them what it means before they draw their own conclusions.

Barclays has spent several years on cloud migration, data cleanup and AI tooling. Putting the CEO's chief of staff in charge of getting people to use it is the right move. Whether it works will show up in usage numbers, and Barclays should start publishing them.