Big banks don't have a data advantage and it helps to be fast.
Small banks and credit unions are operating in an increasingly tight environment. Low money rates continue to hurt margins, fees are getting compressed, and compliance is as burdensome as ever for credit unions growing past asset thresholds. Oh, and those fintechs are always nipping at the heels. While smaller banks and credit unions pride themselves on their close relationships with customers and members, staying a step ahead in the battle for the consumer is going to require a more aggressive investment in new technologies.
One technology that comes up repeatedly in conversations is artificial intelligence (AI) and its subset, machine learning (ML). Machine learning crunches giant data sets from multiple sources to automate and improve decision-making. A 2018 World Economic Forum report, conducted with Deloitte, found that AI/ML are likely to affect every aspect of banking, from customer relations to lending to asset management. The report also noted that, while dominant banks once relied on giant balance sheets to succeed, banks of the future will win by harnessing data through AI and ML to create highly customized products and services.
The problem is that too many smaller financial institutions are unfamiliar with AI/ML.
Some smaller banks and credit unions may stop at this point and worry that they don’t have enough data to compete with the giant banks. Not true at all. They already have the data they need to succeed. One credit union executive told me that, with close to 700,000 members and an even bigger loan history than that, she has plenty of data for robust ML modeling.
The real risk is not acting. According to a 2018 survey by Fannie Mae, 76% of larger banks and 67% of mid-sized banks are familiar with AI and ML technology, compared to only 47% of smaller institutions.
On the customer acquisition side, smart use of AI/ML means a better predictive matching of the right product or service with the right customer. On the customer service side, it can mean automating the credit and debit card management system to speed up the authentication process. A growing number of small banks and credit unions are going even further and deploying AI/ML in the core business of credit underwriting. ML models can crunch through orders of magnitude more data than traditional lending models to improve predictions of borrower risk. Smaller banks and credit unions already have the data they need in loan performance histories or existing subscriptions to the credit bureaus.
This new-found data edge gives smaller banks and credit unions a greater ability to win new customers and members from within underserved communities. Clients who have used Zest AI have increased approval rates by almost fivefold among thin-file or near-prime borrowers. For one mortgage lender, a Zest-built lending model was able to shrink the approval rate gap between white and African American borrowers by 32%.
Small banks are getting help learning more about these important new technologies. The Independent Community Bankers of America is working to educate its members in AI, through a program called ThinkTECH which highlights promising AI fintech partners for smaller banks. As the Fannie Mae report found, bank executives say that the complexity of integrating AI and ML applications with existing bank infrastructure is the biggest impediment to its adoption. Partnering is crucial to speed up adoption.
The potential payoff from thoughtfully scoped and well-coordinated ML projects can be huge: more approvals, reduced risk, faster decisions. Automating easy-to-approve applicants give lenders more time to focus on what banks and credit unions excel at: taking care of customers and members. The Brookings Institution says that banks, insurance companies, and investment management firms could save more than $1 trillion by 2030 thanks to AI technologies, with the bulk of those savings benefiting banks. And not just the big ones, either.