AI Crisis or Opportunity?

For credit unions, the answer is both … and neither

I had breakfast with a credit union CEO at a conference last year who told me they were in the middle of an “AI audit.” I asked what that meant. She said they were figuring out how many positions they could eliminate.

Six months later, a different CEO told me they’d just hired three people because their AI tools had generated enough new loan volume that they couldn’t keep up without them. Same technology, completely different philosophy about what it was for.

That gap is what I keep coming back to, because I hear the first version of that conversation a lot more often than the second. Credit union leadership is under pressure to show that AI investments are paying off, and the easiest way to demonstrate that is through headcount reduction: fewer people doing the same work, which looks clean on a spreadsheet. But chasing that math means you’re using AI to tread water faster. The credit unions that are actually winning have reframed the question entirely, using AI to grow their loan volume and then hiring to keep up with what they’ve built.

Two Ways to Use the Same Technology

You could use AI to cut your way to efficiency. Or you could use it to reach more members, close more loans, and grow your book of business to the point where you need more staff than you had before. The first path produces a leaner institution. The second produces a stronger one. Part of why more credit unions haven’t taken that second path is that growth is harder to promise in a board presentation than savings are. Cutting five positions has a number attached to it. Reaching 3,000 members you weren’t reaching before is a bet, and leadership teams don’t always have the appetite for bets. But that reluctance is exactly what creates the opening for the institutions that do.

The credit unions getting this right are the ones that started with a data foundation. They know which members are likely to need a car loan before those members start shopping. They know where people are dropping off on their websites and what that costs them in closed business. They know which branch interactions convert and which ones don’t, and they’re using that information to make better decisions about staffing, products, and outreach.

Good data tells you what happened, but more importantly it tells you what to do next. The institutions that are growing treat data as the strategy itself, not just the thing that runs underneath it. The ones that bought tools without building that foundation are going to wonder, a few years from now, why the results never came.

Read More on Fintech : Global Fintech Interview with Rob Young, Managing Director – UK at InDebted

The Conductor Problem

A train without a conductor doesn’t run better on its own. It runs off the rails, or it stops. Someone has to keep feeding the engine, monitoring what’s happening, adjusting in real time, and that ongoing human judgment is exactly what makes the whole system function.

AI in your credit union works the same way. The implementations that are actually delivering results have humans behind them, either internal staff who own the strategy and keep it calibrated, or a partner whose job is to manage the system over time. What’s consistently not working is the launch-and-leave approach: deploy a product, celebrate the announcement, and assume the outcomes will follow.

This matters more right now than it probably ever has. Tens of thousands of new technology companies launched in just the first quarter of 2026. Some of them are real businesses with real products. A lot of them are not, and some were essentially built by AI, pitched by AI-generated marketing, with no one particularly experienced minding the engine. Telling the difference has become one of the more important skills your leadership team needs to develop. When you’re evaluating a vendor, the pitch deck tells you what the product does on the day you buy it. What you really need to understand is who’s running it a year later, and whether the company behind it is built to last.

The Real Question

A recent Filene Research Institute survey of credit union leaders found that most credit union leaders still consider themselves in the early stages of AI adoption. That tracks with what I’m seeing: most institutions are still trying to answer the ‘what do we do with this’ question, which means the gap between early movers and everyone else is still very closable.

There’s a version of the AI conversation that sounds like strategy but is actually just anxiety. How do we leverage AI? What do we do with it? Where does it take us? Those questions aren’t bad, but left unanswered they tend to drift toward whatever’s easiest to measure, and what’s easiest to measure is usually cost, not growth.

Credit union executives are right to feel some urgency. There’s a real window here, and it won’t stay open. But urgency pointed in the wrong direction produces its own losses, ones that just don’t show up on the same spreadsheet. The members you didn’t reach, the loans that went to a competitor, the digital experience that quietly sent someone elsewhere — none of that gets a line item.

The institutions that come out of this era in better shape than they entered it will be the ones that decided, early enough, that AI was a growth tool first. Most boardrooms haven’t framed it that way yet. They’re still asking how to get leaner, when the more useful question is how to get bigger — more members, more products per member, more of the lending business that’s currently going somewhere else.

That shift in thinking doesn’t take a new vendor or a bigger budget. It just takes an honest conversation about what you’re actually trying to build. The credit unions I see pulling away from the pack right now made that call a year or two ago, quietly, without a lot of fanfare.

About Appli

Appli was created to personalize digital banking and web experiences by bringing smart interactions directly into your lending and deposit pages.