You Can’t Spell Pain Without ‘AI’

PUBLISHED MARCH 2024

Imagine this: A new tech tool promises to help your business improve the customer experience, bolster security, increase staff efficiency, lower your costs, and reduce compliance burdens.

What’s not to like? If those don’t sound like things that keep you up at night, you’re probably not a banker.

The tech, of course, is artificial intelligence, which is on the lips and minds of most every forward-thinking executive on the planet: How to use it, how to maximize its capabilities, how to turn that cost center into a revenue center. For bankers generally, there’s an additional “how to” at work: How to ensure data security? And for community bankers in general, it’s “how are we going to pay for this?”

AI, banking executives say, has already proven its ability as a differentiator, and 75 percent of the nation’s banks are using some form of it. But as with the disproportionate toll exacted on smaller banks by one-size-fits-all regulatory demands crafted to rein in big banks, some serious discussions will be in order. 

Their numbers are not inconsequential. As of Dec. 31, the FDIC had in its database 4,590 banks and all but about 1,000 of them had assets below $1 billion—a rough dividing line between “small” banks and the mids and majors. Executives at those smaller institutions are trying to balance the potential benefits of AI with the likely costs, and they’re under some pressure to get the calculus right before the loss of market share sets in.

Those are the two buckets to which Jackson Hataway, CEO of the Missouri Bankers Association, divides the issues. And security is a big one.

“When you’re talking about data in banking, it is sacrosanct,” he says. “It’s the most privileged information that people have. Because of that, banks will be slower to adopt some of the AI, regardless of size or scale. Some large banks are already deploying it, but we’re starting to see more community banks with the same opportunities for purposes of simplifying customer interactions.”

Anybody doing online banking today has likely run into the intrusive chatbot inquiries, and “consumers can get frustrated with them as they are currently developed,” Hataway says. “AI can make that more seamless, help manage accounts better, provide more alerts and more timely alerts. There are a lot of positive uses, but slow adaptation as banks protect their data set in-house while they explore ways to leverage this.”

At Academy Bank, the lever is already being pulled, says COO Tom Kientz—the bank’s AI chatbot already handles more than 20,000 interactions every month. 

He calls the introduction of that technology “one of the easiest examples of enhancing the client experience, because customers can engage with the bank at any time, day or night. Our bot answers questions about locations, mortgage rates, business loans, and more,” he says. “And the best part, the bot continues to learn.”

That’s the outward focus. The inward counterpart is where things get more interesting as an employer. Someone is going to have to, in effect, manage the tech to ensure it’s doing what it’s supposed to.

“IT related positions will increase, but not as fast as you may think,” Kientz says. “We will need more developers, programmers, data scientists, etc., but the bigger lift will come from re-skilling our existing workforce. AI, machine learning and RPA will greatly reduce the routine, mundane tasks of today but our associates will need to hone their skills around critical thinking, problem-solving, and creativity.”

Security, says Hataway, will play an outsized role in adoption.

“What happens as AI becomes more powerful,” he says, “is that there are more tools connected to AI. What does that mean in terms of fraud rates? There’s a lot of concern by the security industry to insure against ways bad actors might try to use AI. There’s a lot of concern about what happens with fraud. So a lot of investment and energy is going into preventive measures.”

There will be, he says, national councils weighing in, likely intervention from Congress, and forays by the FDIC and other regulatory bodies. “Everybody will want to put controls on the potential damage,” he says.

One needs look no further than the impact of big-bank-focused regulation to see where that’s going. After the Dodd-Frank Act became federal law in 2010, it sharply raised the compliance costs across the industry, but smaller banks lacked the proportionate resources to play at the level of bigger banks.

Less than 15 years later, nearly 2,000 of the banks operating that year no longer exist. The vast majority of those were rural and community banks. Some have simply shut down operations; most were absorbed into larger institutions. There’s no loss of irony in pointing out that legislation designed to rein in big banks had the effect of making them … bigger.

The question now is, will the same be true of AI and the efforts outside the banking sector to determine how it can be used among financial institutions. 

Whatever impact is coming, Hataway doesn’t think it will be dramatic—at least not at the outset. 

“I’m probably a bit of an outlier on this, but at least in banking, you will see such specific-use cases and such careful moves forward, primarily because of security concerns and protected data sets,” he said, anticipating a moderate embrace of adoption. “For the most part, I see more incremental gains as AI is deployed in financial services rather than major leaps and bounds. There will be opportunities to do more with it, but given the way the regulatory environment is looking at these advances, I don’t know that it’s possible to do anything too advanced without having every regulatory agency checking all the boxes, and that will have implications on what banks can and will do.”