AI isn't the bogeyman credit unions fear it to be

Register now

When VyStar Credit Union opened its doors in 1952, few could imagine the waves of technological change that would transform the credit union industry. There was a time when high-tech meant drive-through windows, mainframes, and later advances such as touch-screen ATMs and mobile check deposits. Now we’re all talking about artificial intelligence, big data and cloud computing. But even though the technology has changed, the reason for embracing it hasn’t: faster and better service that strengthens our member relationships. That’s been at the heart of our business for years no matter what technology we use.

Nonetheless, most credit unions have been slow to embrace this latest wave of AI. Mention the words “machine learning,” and many credit union colleagues see visions – perhaps even nightmares – of robots replacing people or going rogue with ruinous lending decisions. Others worry that our data sets may not be robust enough, especially compared to the vast repositories of much larger financial services firms. Still, others are skeptical that these investments will ever pay off. Small wonder that only 5% of credit unions have incorporated AI/ML solutions into their lending business, according to a study of a broad array of mortgage lenders released last fall by Fannie Mae.

At VyStar, we believe that we can’t afford to wait. We’re conservative by nature but that doesn’t mean we can’t also be at the forefront of technology adoption and institutional change. That’s why we are proud to be one of the first credit unions to adopt AI-powered underwriting. Last week we announced a software partnership with a leader in this space. Sure, we look forward to seeing some cost-savings associated with greater automation, but what really excites us is the opportunity to offer instant decisions, better pricing, and personalized service to our 675,000 members.

Like most credit unions, we rely on a loan operating system with certain traditional rules of thumb – and then, once or twice a year, convene our lending and credit policy teams to review and manually recalibrate our credit decisioning models as market conditions change. That system was pretty good at handling member loan applications that were clear approvals or rejections but not as effective at sorting out those cases that could go either way.

Using an AI-powered underwriting model, we can take advantage of vastly more data, more sophisticated math, and increased automation to more nimbly respond to the market – and more precisely evaluate and price the risk of new or existing members seeking a loan. Right now, it takes between 15 minutes to an hour for us to evaluate an applicant. With AI-powered underwriting, we can do it almost instantaneously – allowing us to deliver a superior member experience while closing more loans. If we can say yes to more members on the spot, our employees can avoid the prospect of a lengthy process that leaves the applicant upset or disappointed and have more opportunities to cross-sell them with other relevant lending products.

The AI will also help us flag those applications on the margins so that our underwriting professionals can take a closer look and engage those members in a conversation. Rather than robots replacing jobs, I think AI-powered underwriting will make our loan officers happier about being able to spend more time with our members to figure out a better way to help them close that mortgage or car loan.

AI-driven underwriting is powered by complex algorithms and big data sets, but it is really about getting back to the old days of human employees getting to know our members better. A more custom machine learning model, with automated tools that let us tune it more than once a year, frees us from the constraints of off-the-shelf loan origination systems so we can deliver more personalized, white-glove customer experience and guide the pace of future innovation. In short, it’s about what has always distinguished VyStar and thousands of other credit unions from other financial institutions: a relentless focus on our members.

For reprint and licensing requests for this article, click here.
Artificial intelligence Machine learning Cognitive computing Technology Lending Florida