When Bank of Ireland goes looking for its first-ever Chief AI Officer, the hire tells you something about where retail banking thinks its next decade will be won. The lender has named Prag Sharma to the role, poaching a senior figure from Citigroup and handing him a seat that did not exist at the bank a year ago.
Who the bank hired
Sharma takes up the post in October. At Citigroup he built and led the Global AI Centre of Excellence, an operation that shipped AI-powered products across 96 countries, and most recently he worked as a director in Citi Institute’s Future of Finance think-tank. His background runs deeper than one employer. He has held innovation and emerging-technology roles at Deloitte’s advanced analytics practice, University College Dublin’s Insight Centre for Data Analytics, and Accenture’s integrated markets group. He earned a bachelor’s degree in electronic engineering, a PhD in computer vision and pattern recognition, and an MBA, all from UCD.
That mix matters. Sharma is not a management consultant who picked up AI as a buzzword. He is an engineer who has spent two decades building the systems, then moved close enough to strategy to know what a bank board actually wants from them.
A role that signals intent
Creating a C-suite position is a statement, not a formality. Ciarán Coyle, Bank of Ireland’s group chief operating officer, said AI sits at the centre of the lender’s updated strategy through 2028. “With a proven track record of delivering impactful AI globally, Prag brings the vision, leadership and expertise required to safely accelerate our AI transformation and unlock meaningful value across the group,” Coyle said. He added that the appointment reflects the bank’s ambition to be “at the forefront of responsible AI innovation and adoption.”
Note the word Coyle used twice in spirit if not in letter: responsible. Banks operate under regulators who care a great deal about how automated decisions get made, who they affect, and whether anyone can explain them after the fact. Putting a single senior executive in charge of that transformation is partly about speed and partly about having one throat to hold when a supervisor asks how a credit model reached its conclusion.
Why banks are minting AI chiefs
The appointment fits a pattern spreading across the industry. Lenders have moved past the phase of scattered pilots and proofs of concept, and many have concluded that coordinating AI across fraud detection, credit risk, customer service and back-office automation needs an owner with real authority. Fragmented efforts produce fragmented results, duplicated spending, and models that nobody governs consistently. A chief AI officer is the organisational answer to that problem.
The technology has genuine pull on both sides of the ledger. On defence, AI systems catch fraud patterns and flag suspicious transactions faster than rule-based tools ever did. On growth, they promise sharper credit decisions, more personalised products, and cheaper service through automation. For a bank the size of Bank of Ireland, even modest gains in those areas translate into meaningful money, which is exactly the “value across the group” Coyle referred to.
The harder part comes next
Naming an executive is the easy step. The difficult work is everything that follows, and Sharma will inherit the same tensions every bank faces when it scales AI. Models have to be explainable enough to satisfy regulators. Customer data has to be handled under strict privacy rules. Automated decisions about lending carry fair-treatment obligations that a black-box system can quietly violate. And the reputational cost of getting any of that wrong lands on the bank, not the vendor that supplied the model.
Sharma’s Citigroup experience across nearly a hundred markets is presumably why Bank of Ireland believes he can navigate those constraints. Building AI that works is one skill. Building AI that works inside a regulated institution without tripping a compliance wire is a different and rarer one.
The real measure of this hire will not be the title or the press release. It will be whether, by 2028, customers notice their bank got faster and smarter without anyone noticing it got less fair. If Sharma pulls that off, expect every other lender still without a dedicated AI chief to start writing the job description. For more coverage of AI in finance, visit Mylistingo.
