3 ways Deutsche Bank drives AI adoption

Financial institutions are implementing AI throughout their organizations at a steady pace. However, even with investments made and use cases identified, AI can only accomplish so much without users on board.

“If the businesses [within a bank] don’t care to use it, it’s just not going to change anything,” Inwha Huh, managing director at $1.4 trillion Deutsche Bank, said last month at Sibos.

At Sibos 2023, from left, are: Shelby Austin, Grace Lee, Mike Hughes and Inwha Huh. Photo by Whitney McDonald.

Decision-makers at other banks agree.

Mike Hughes, global head of custody product development at Citigroup, said: “If you don’t get user adoption [of AI], it’s just a huge waste of time, effort and energy.

Similarly, Scotiabank Chief Data and Analytics Officer Grace Lee cautioned that if AI isn’t integrated into a bank’s infrastructure and culture, “Then we’re going to continue to under-invest. … And that’s a recipe for us to spend another couple of decades without AI fundamentally changing the way we live and work.”

To make the transition to an AI-driven institution, Deutsche Bank has identified three ways to drive the technology in its operations:

1. Pushing adoption: Bank employees who understand how AI will change their day-to-day work lives will want to use the technology.

“The No. 1 thing that comes to mind for me is people,” Huh said. Employees much think differently about AI in order to understand and use it, therefore any adoption effort must go beyond senior management and into middle management and users to be successful.

2. Adjusting the operating model: Banks must undergo fundamental changes to their operating model to benefit from AI, she said. For example, AI must be part of the business structure and client experience from within, rather than an added capability.

“It’s not enough to stick in cool new technology,” Huh said.

3. Reengineering processes: Banks often run proofs of concept that don’t lead to change, however this cannot be the case with AI.

“The process [of] reengineering up front as you embed new and great AI tooling is also critical,” Huh said. “And unless [we] reengineer the way we do stuff. … nothing’s really going to change.”