The AI revolution in credit union land: Starting small is smart

It was at an intimate gathering of credit union executives and fintechs in June where a new reality hit me: credit unions suddenly had realized that  AI—artificial intelligence tools such as ChatGPT—are going to transform the industry at a speed that would have seemed unimaginable.

And the transformations will be huge and profound.

But it is all starting with little steps. That’s key.

Two credit unions told me about their experiences with AI and why they are taking small steps.

Case in point: One Nevada Credit Union, a $1.38 billion institution, started its AI journey by taking all of its retail procedures and policies and dumping them into an AI powered database, Steve O’Donnell, an executive vice president at the credit union, related to me in a podcast. Why? Historically in credit unions much information is siloed and each department had created its own manual. So an employee working in lending might have a different answer to a member question than an employee working in collections or a teller. All three might be right, as far as their information went, but what they all missed was an institution-wide perspective.

And members sometimes were frustrated that in three different conversations with credit union employees they got three different answers.

“What we’re doing now will really accelerate how we deal with member service,” said O’Donnell.

A plus, said O’Donnell, is that this AI powered member service is actually saving One Nevada money. How? Historically a member service rep would field a member call and it might require the rep to call an expert in, say, lending. Such calls typically took maybe 15 to 20 minutes to get an answer that satisfied the member. But using this AI powered database has dropped the average call time to three minutes, said O’Donnell. That delivers powerful cost savings—and also much happier members who get the info they need faster and without long waits on hold.

Step back a moment. Say “AI” and many of us immediately think of The Matrix where brainy computers rule and humans are dispensable. Nobody I have talked to in credit union land—and I have talked to many CU executives—has that kind of sweeping, grand view of what AI will do in credit unions, at least not in the near term. Long-term, AI likely will have a transformative impact as big as the advent of the Web in the mid 1990s and the debut of the smartphone circa 2010. But, for now, baby steps rule.

Up in Washington State, David Eldred, chief experience officer at billion dollar Solarity Credit Union said that, for him, a lightbulb went off in his head that illuminated a path forward with AI when he realized he could use ChatGPT to create custom tools to work inside his credit union.

Let’s back up a step. Eldred had had a conversation with a high powered consulting group that had been working on various AI tools and it all sounded interesting until Eldred asked what it would cost to install the tools into Solarity. The answer was $500,000 to $1 million.

That’s when Eldred decided to start slow and with a small price tag. It started by asking AI to develop interview questions for an executive guest on the Solarity podcast and to outline some new content for Solarity’s web page. “What used to take three days now takes us 30 minutes,” Eldred told me in a podcast.

Then he got bolder. Eldred has a practice of meeting individually each week with his staff where they go over measures of success. “Our methodology is to take notes and to post the key takeaways and action items to our Yammer page.” It used to take Eldred 30 minutes to type up the notes for each meeting.

Now he lets Microsoft’s AI tool, Copilot, do the note taking: ”What used to take me a half hour is available 30 seconds after the meeting ends.”

He added: “I’m saving three or four hours a week that can be put to more productive uses.”

Of course Eldred has bigger, bolder plans for deploying AI at Solarity. O’Donnell does too at One Nevada. But both say they are sticking with a go slow, stay small approach with AI for now because they are getting meaningful results that are helping build confidence in the utility of AI to a credit union.

Here’s the big question: what impact will AI have on employees? Right now, many employees fear their jobs will be eliminated by AI and those fears aren’t entirely off-base. A lot of routine labor stands to be done better and cheaper by machines and that transformation is already starting in many credit unions.

Will jobs be eliminated?  Said O’Donnell: “I think credit unions are going to have hard decisions to make. For us, it’s really about how do we up-skill our employees and get them to that next phase of how work is changing so that we can keep going on.”

Want a clue about the coming changes? I just asked Google’s Gemini AI tool this: “How will credit union staffing change as AI tools are deployed?”

Do similar and you may be encouraged.