JPMorgan Chase poised to take gen AI from experiment to implementation

Banks’ use of generative artificial intelligence is on the cusp of shifting from an experimentation phase to implementation, according to the chief data and analytics officer at JPMorgan Chase .

Next year could mark the start of a transformation in banking, with “a little less experimentation and more rubber hitting the road,” according to Theresa Heitsenrether, who spoke Tuesday afternoon at American Banker’s Most Powerful Women in Banking Conference in New York City.

“We’re starting to see an evolution of [AI] teams that are multi-faceted,” Heitsenrether said. “I think we’ll see more of that, and that will drive more traction” around substantive change.

What’s ahead for banks in AI was the topic of the first panel at this week’s Most Powerful Women in Banking event, which celebrates the top female executives in banking and finance. Heitsenrether, who was promoted to her current role in 2023, was joined on stage by Christine Livingston, a managing director at Protiviti who works with banks on their AI strategies.

Large banks have been in an AI arms race for more than a year. A recent analysis by Evident that gauged banks’ efforts to develop and deploy AI technology put JPMorgan in the top spot, followed by Royal Bank of Canada and Citigroup. Rounding out the top 10 were UBS Group, Wells Fargo, TD Bank, ING Group, Bank of America, BNP Paribas and Morgan Stanley.

The Evident AI Index scores the 23 largest banks in North America and Europe, using public information such as talent, innovation and leadership. It will soon cover the top 50 banks.

AI is not a new strategy or tool at JPMorgan, the largest U.S. bank by assets. When generative AI came onto the scene, “it opened up a new aperture,” Heitsenrether told the audience, which was a mix of bankers, regulators and some vendors. About 195,000 JPMorgan employees are using the bank’s generative AI portal in some fashion, including contract reviews and communication, she said.

Of those 195,000, about 60,000 are “really active users,” Heitsenrether said.

“It just makes people more efficient,” she said. “Getting people’s hands on the tools is sparking a lot of ideas … and once you see your colleagues doing it, it sparks some curiosity.”

Banks, many of which are still heavily dependent on manual processes, continue to figure out ways in which they can effectively employ AI, Livingston said. At the same time, they’re balancing potential use cases with the costs and inherent risks of using AI tools, she noted.

For banks, “thinking about risk is not a new concept,” Livingston said. She said she reminds those with whom she works to “lean on concepts and frameworks” that they’ve already built.

Heitsenrether suggested that banks remember that AI is one option among many.

“Generative AI is not going to be the tool of choice for everything,” she said. “Sometimes you may need a more traditional method. This is just one tool in the box.”