By Evan Sparks and Khalil Garriott
What are the bank business models of the future? And can AI ensure a frictionless world of core conversions?
ABA’s Core Platforms Committee, which launched in 2018, is currently examining these questions through exploring trends in artificial intelligence, digital asset growth, yield-seeking behavior and the general competitive landscape. The end goal: to identify what’s next for bank strategy.
One hypothesis: Net interest margins will get compressed, says past committee chair Kristiane Koontz, EVP for treasury management and payments at Zions Bancorporation.
“Consumers will have more visibility into yield options, driven by AI tools that allow them to get that information and actually move the money around,” she says. (This trend would be compounded if stablecoin providers were to continue to be able to use the Genius Act loophole to pay yield-like rewards.)
In that future landscape, Koontz says, “I have my own agent that’s my personal CFO, and I care less about relationships and more about maximizing yield.”
Business accounts will have the same capabilities, and they can also use AI agents to manage cash flow, decreasing the amount of idle cash parked in operating accounts and further reducing the float that banks access through their deposit accounts.
How do banks respond? Core Platform Committee members see two potential game plans. First, more banks pivoting to becoming infrastructure providers — an expansion of banks’ role in banking as a service, or in rolling out tokenized deposits and other digital asset platforms. Another approach involves banks needing to differentiate and find stable ground to compete.
The latter approach involves specializing in particular customer segments. Once this segment is identified — Koontz throws out the hypothetical example of Dog Bank, a bank for dog owners — the bank then begins designing products with bundled and tailored offerings that will provide customers with fewer incentives to go price shopping or switch. It’s the same model that hotels and airlines use; loyalty programs offer benefits for frequent travelers to persuade those high-spending customers not to shop around for other lodging or airfares.
In Koontz’s Dog Bank example, the bank might offer a doggy daycare amenity for high-value customers in underutilized bank real estate, increasing loyalty and the cost of switching.
“To do that, I’ve got to have really good data and hyper-personalization, and my core has to support this,” she says. “This is a silly example, but it is meant to inspire some thinking on what more tailored, bundled products and business models might look like.”
Therein lies the rub. Vendor management is limited, and the industry, by and large, still has a fear of core conversions.
“The legacy technology does not allow us to quickly spin up new products,” says Clay Adams, president and CEO of Mascoma Bank in New Hampshire and future committee chair. On most legacy platforms, this means banks can offer products under the “Henry Ford strategy: ‘whatever color they want, as long as it’s black.’”
The legacy cores are starting to move faster, and core contracts will start to shorten. Key to the transformation is data — data that banks have always had, Adams says, but just couldn’t access. Data enables a “more micro-segmented approach.”
To execute this transformation, Mascoma Bank expects a year-end switch to a headless core, from Thought Machine, that will enable the bank to create products using “smart contracts” much quicker than they can today. That will allow them to serve their customers’ needs more readily, Adams says, “helping them realize opportunities that, prior to now, we couldn’t see for them, to get them in the right products and services.”
Other critical technologies for this future: real-time cores to support the demand for real-time payments, and access to APIs for banking services.
One bank represented on the core committee recently created a virtual AI assistant that does Gen AI better than the core itself does Gen AI. The latest ABA survey data shows that 8% of banks aren’t using AI at all.
“Banks will have to make a choice,” Koontz says. “Do I let my customer consume my banking services in the user interface of an AI agent, or in the user interface, in a digital banking layer that I control?” This leads to another key tech need: upgraded identity and access management systems, because identity will go from “know your customer” to “know your agent.”
In the future, the core might be part of the tech stack, but bank leaders believe it won’t be the center of the tech stack. The “bank of the future” will know how to design the best products, and the “core of the future” will enable that aspiration.
“The future is rewriting itself,” says Julieann Thurlow, Reading Cooperative Bank CEO.
Evan Sparks is editor-in-chief of the ABA Banking Journal, and Khalil Garriott is executive editor of the ABA Banking Journal.
