Can banks win at AI?

Enjoy complimentary access to top ideas and insights — selected by our editors.

Does advanced artificial intelligence provide an advantage to banks or to their fintech competitors?

This question came up during JPMorgan Chase's third-quarter earnings call last week.

Longtime bank analyst Michael Mayo, managing director at Wells Fargo Securities, questioned JPMorgan Chase's increased spending on technology, including on AI. The bank's expenses of $8.5 billion were up 7% year on year, largely driven by continued investments in staffing, primarily in front office and technology.

"As it relates to AI specifically, which is the talk of the town, the consensus among people outside the banking industry is that banks will not win that battle, including JPMorgan," Mayo said. "You won't control the front end. What are you doing with AI to make a difference now? Or is this simply a moonshot?"

Jamie Dimon, chairman and CEO of JPMorgan Chase, said he disagreed with the statement.

"AI is an extraordinarily good tool to use," Dimon said. "We just put a woman who is running it at our table," he said, referring to Teresa Heitsenrether, who was appointed the firm's chief data and analytics officer and a member of JPMorgan Chase's Operating Committee in June.

JPMorgan Chase uses AI for risk, fraud, marketing and prospecting, Dimon said. 

"And the management team is getting better and better at saying, how can we use data to do a better job to reduce errors? To serve clients better? To have a salesperson have copilots, that they know why even the client's calling or something like that," Dimon said. "And so we simply have to do it. Does it create opportunity for disrupters to come in? Yes, of course. That's always the truth of technology and — but we'll be quite good at it."

In a follow-up email exchange, Mayo explained that advanced AI is likely to intensify the competition between big banks, small banks, fintechs and pure technology companies.

"Our theme on the banking industry is 'Goliath is winning,' implying that the largest banks should have an advantage in using AI given [they have] so much more proprietary data," Mayo said. "However, the counterpoint is that certain smaller banks have more nimble tech stacks, and that fintech and big tech have more agility to better manipulate and deliver AI-enabled solutions." 

Large banks are in one corner with their vast amount of proprietary data that can help customize solutions customer by customer, he said. In the other corner are technology companies that don't have the burden of older infrastructures that make extracting data so difficult. 

"The middle ground is the potential for smaller banks, fintechs and others to piece together solutions that use modern technology and proprietary data from a variety of sources," Mayo said. "I don't expect this debate to be resolved anytime soon."

Some industry experts expect banks to win this AI battle.

"At Finovate Fall this past month, I saw several fintechs offering ChatGPT-like front ends for banks and credit unions to adopt, and a large bank like JPMorgan Chase is certainly capable of building its own version if it chooses to," said Aaron McPherson, principal at AFM Consulting. 

Generative AI will help disrupters trying to create a new way of banking, acknowledged Alex Jiménez, managing principal, financial services consulting at EPAM. 

"But just as startups can do it, so can banks of any size," he argued. "For the past decade, we have been hearing that neobanks are disrupting banking and will put banks out of business because they provide better customer experiences. While there is some truth in that, the reality is that most banks have improved their digital experience."

Large banks like JPMorgan Chase, Bank of America and CapitalOne, and smaller banks like Frost, Coastal Bank and Grasshopper provide digital experiences that compare well to challenger banks like Chime and Varo, Jiménez said.

"I expect that banks will deploy generative AI on their own, with partners like EPAM, or with the help of their digital providers like Backbase, Q2 and Alkami to compete with any entrants," Jiménez said.

Artificial intelligence

How banks can cross the 'uncanny valley' as AI becomes more human

Banks need to — and can — take steps to manage the phenomenon, first defined by a Japanese academic in the 1970s, where artificial intelligence looks, feels or sounds human and gives customers a creepy feeling, experts say.

Many banks are already using traditional forms of AI, McPherson noted. For instance, Bank of America's Erica virtual assistant, which recently surpassed 1.5 billion customer interactions, uses natural language processing and machine learning.

"There are already big banks that are leading in this area, and what we are seeing is a democratization of the technology so that it can be packaged and sold to financial institutions that do not have the resources to develop their own version of it," McPherson said.

Such chatbots are a complementary channel to telephone, branch, mobile and web interfaces, and not a replacement for them, he added.  

"So I'm not even clear what 'winning' means in this case," he said. "We are talking about a technology, not a business model."