AI 'Enhancing' Credit Unions' Visions

This could be the most simple, real-world explanation of artificial intelligence: Type a word in the Google search bar and Google automatically fills in what people are looking for.

In more detailed terms, an AI solution compiles massive amounts of data to scale with incredible speed, while eliminating much of the bias that humans intentionally or unintentionally carry around with them.

As Scienaptic AI EVP of Client Impact Eric Steinhoff explained during a credit union/AI panel discussion, “Sometimes humans call in sick or humans quit or sometimes are hungover and in a bad mood and might be like, ‘I don’t feel like doing this today.’ Computers have other challenges, but don’t bring that human baggage.”

For credit unions that use or are beginning to use AI tools and programs, it’s not about using AI to replace employees.

“Our vision is to more rapidly approve loans to our members and free up time for our loan officers and underwriters to work with members who aren’t automatically approved to get those people approved through whatever means it might be. Speed and then also being able to grant more loans to more members more often is our vision of what we want to see happen with AI,” said Floyd Rummel, CEO of the Sturgis, S.D.-based Northern Hills Federal Credit Union ($127 million in assets, 9,955 members).

The chief technology officer of the Denver-based Credit Union of Colorado ($2.2 billion in assets, 148,078 members), Mark Rowan, echoed Rummel’s assessment. “I see AI as not a replacement for staff, but to help augment staff resources. So if we can offload a lot of the heavy lifting to machines, then it allows our staff to do other things like member experience and stuff that credit unions do extremely well,” Rowan said.

Rummel’s credit union began its journey with AI only weeks ago with an investment in the Scienaptic AI platform. The long-time credit union executive said the time for AI has come because members no longer are tolerant of the time it takes a credit union to decide their financial future. “Times have changed and our membership demographics have changed; and our younger members, they don’t have the patience that their parents and grandparents had with the credit union,” he said. “And if we’re not in a position to respond to those members immediately, they will find a financial institution that does and there’s no loyalty, so they’re not afraid to jump ship.”

While AI has been around for years, growth in the AI industry appears to be hitting its stride. The Stamford, Conn.-based technology research company Gartner stated AI software revenue is forecast to hit $62.5 billion globally in 2022, an increase of more than 21% from 2021.

A study by Autonomous Research predicted AI will reduce operational costs for financial institutions by 22% in the next decade.

Calculating the true cost of AI for credit unions is tricky as the range could be anywhere from $6,000 to $300,000 or more for the AI software, depending on the solution needed.

After a month of using AI for lending applications, Alicia Carlson, who works in credit systems – consumer lending at the Richland, Wash.-based Gesa Credit Union ($4.8 billion in assets, 268,304 members), said it’s too early to tell how much money the credit union has saved with AI so far. For auto loans, she said, “Right away we saw actually a big lift in those system approvals.” She noted that approval rates went from 15% in the first week of AI approvals to 35% by the second week.

During a recent panel discussion hosted by Neuton.AI, Todd Lindermann, SVP at the Santa Rosa, Calif.-based Redwood Credit Union ($7.1 billion in assets, 312,084 members), said his credit union has used AI for some time as a way to look beyond simply applying it to lending efforts, but more with predictive analytics as it relates to risk in nearly all credit union product areas. “So we use AI to basically download all of our transactions. We’ll look at that and try and predict the likelihood that someone would need higher cash limits. And we’re going to expand that over to POS limits, basically on all cards and limits kind of across the board, and try and use data to predict what they actually need,” he said.

As many AI-using credit union executives see it, AI has provided a new level of cooperative service for members. “We all want to serve more members more quickly, and for me, for us, AI is really just enhancing our vision,” said Daniel Estal, vice president of lending for the Watertown, N.Y.-based Northern Credit Union ($430.7 million in assets, 35,836 members).

Estal said his team sat down to discuss what the vision is for the credit union and how AI could boost that vision. “It’s going to help take our data and everything that we’ve already kind of compiled and know of our membership, and just help us make better decisions on how those members perform – and that’s huge!”

He said AI has the ability to look beyond just the FICO score and other credit-reporting numbers. AI can compact so much data so quickly that it has been a big relief for what was the everyday work load for his team. “The benefits of AI … it’s bringing all of that scale into the equation quick and taking that data crunch off of our human minds and just giving us the answer: ‘Will this person pay us back or not? Have they in the past or not?’ And really help us help more members more quickly.”

Neuton.AI Chief Technology Officer Blair Newman said the chaos of the pandemic appears to have caused an awakening for some credit unions as to the need for using AI. “The pandemic has maybe kind of accelerated things where, to some degree, there has been a paradigm shift to another degree. You need to begin to look at things with a different lens,” Newman said.

PSCU Senior Innovation Strategist Jay Lauer said that new lens involved innovation. Innovation, as in adopting AI as a new way of doing business.

“So credit unions typically don’t have the risk appetite or the financial war chest needed to be too cutting edge,” Lauer said. “And they’re typically going to fall into those fast follower or early majority phases of innovation adoption.”

Lauer said AI adoption by credit unions has gone up, as credit unions have taken the adoption slowly and one step at a time, instead of going all-in immediately. “It provides them the opportunity to survey the environment a little bit, to identify new technologies or new approaches. Then [credit unions] determine if those technologies and approaches are going to benefit the credit union and their membership,” he said.

According to a 2020 survey by accounting software fintech Tipalti, 60% of credit unions surveyed believe a fintech partnership will be crucial to their operations by 2025.

Rummel echoed the value of a fintech partnership and said there’s a lot of work to be done in order to ensure a credit union’s core systems can work with the fintech/AI vendor. Despite that work and even if the credit union is smaller, Rummel encouraged credit unions to take a hard look at AI.

“This is what the future is going to be,” he said.