AI could save America’s community banks — if they allow it

Yerbol Orynbayev is president at TurmaFinTech, a startup focused on helping U.S. banks and credit unions beat customer churn through custom customer data platforms. Prior to his consultancy career, Orynbayev served as the deputy prime minister of Kazakhstan from 2007 to 2013 and aide to the president on economic policy from 2013 to 2015. 

Since PayPal burst onto the scene in the early 2000s, technology has continued to reshape the entire US banking ecosystem. It has posed challenges for regional and community banks that have at times felt existential, but it’s not all doom and gloom. The rapid development of AI opens new opportunities for smaller institutions to slow customer churn and grow exponentially — but they must seize them.

There’s no doubt that the road to technological advancement has been paved with challenges for smaller banks — especially when it comes to the fragmentation of the banking landscape. Consumers now turn to a different provider to make payments than they do to build their savings. They reach for one app to monitor their credit score and another to take out a short-term loan.

With the number of providers and products steadily increasing, it’s no wonder that customer churn rates have crept up. According to J.D. Power, 8% of retail bank customers changed their primary bank from 2022 to 2024, and 13% indicated they were likely to switch in the coming year.

For smaller institutions that rely heavily on strong customer relationships and close proximity to the communities they serve, such a hit to customer loyalty could threaten their survival.

Turning the tide on customer churn

There are reasons to be optimistic. Technological advancement does bring exciting opportunities, and AI offers more than most. Whether it’s by leaning on LLMs to provide better customer service or automating processes to deliver loan approval decisions within minutes, the tech can help smaller banks rebuild the customer loyalty so crucial to their success. But there’s one small hurdle to adoption.

Although 90% of community banks want to initiate digital transformations, they are more hesitant about AI, deterred by a lack of understanding and tight budgets. In 2024, only 40% of community banks said they were incorporating AI into their strategies, leaving many without the AI-fueled capabilities they need to stay competitive.

AI adoption to fit any budget

While smaller banks don’t have the same financial firepower as some of their larger counterparts, they can and should implement AI on a scale that works for them. The way to do that is by targeting their customer data — one of their most valuable assets.

It all comes down to being able to manage, analyze and monetize this data. AI can sift through data to spot behavior patterns — helping banks know when a customer needs support, a new product, or may be at risk of default.

They offer opportunities to upsell, improve risk management, and strengthen relationships by better supporting and identifying at-risk customers.

And there are multiple pathways they can take to gain these capabilities, too. They could go down the cloud platform route — leaning on an external platform’s infrastructure but building their own algorithms — or they could turn to vendor partnerships that provide ready-made AI tools that fit like a glove. If banks want light-touch AI that still helps them build customer loyalty, they have API-based deployment as an option — it lets them integrate external AI models, from providers such as OpenAI or Zest AI, into their own systems.

Proof points are already out there

We’ve already seen impressive success stories across the sector. By the end of 2022, Regions Bank had rolled out ten data products, including an AI model that flags early attrition risk and monitors credit exposure, and a customer feedback analysis tool that has helped save over $1 million a year while identifying customer issues three times faster.

In a similar vein, FVCBancorp has shown that AI has a role to play at the community banking level. Its model automatically accepts or declines applications, streamlining risk analysis and helping approved applicants receive funds within 48 hours.

There are already clear use cases showing that AI helps banks build customer loyalty. Yet we’re still not seeing widespread adoption.

Of course, implementing AI poses real challenges — tight data regulations, hallucination risks, and difficulty finding skilled talent to manage models — concerns that have understandably held some banks back. Still, our understanding of the technology is improving every day, and I firmly believe the benefits far outweigh the risks. In fact, before long, AI won’t just be a nice-to-have — it will be essential to doing business.

How AI can future-proof community banks

Let’s face it: the in-person, in-branch experience is not as relevant as it once was. In the first quarter of 2025, 320 bank branches were marked for closure — they’re quickly disappearing from our towns and cities.

Given that around 73% of global bank interactions now happen through digital channels, it’s no surprise that brick-and-mortar branches are shutting down. This should be a wake-up call: banks must invest in tech-enabled ways to build customer relationships — or risk undermining the very loyalty they are built on.

Technology may have permanently reshaped the U.S. banking ecosystem, but it has also introduced new tools — powerful weapons regional and community banks can use to take on the competition and reduce high churn rates.

The time has come for community banks to get the ball rolling on AI implementation, whether it’s by hiring their own data scientists or working with vendors to explore tailored AI solutions. The fact is, there are affordable, customizable options out there. Banks can choose how far into AI they want to dive — the important thing is that they get the wheels in motion now.

The brick-and-mortar era is fading away, and the digital age is within touching distance. Smaller banks just need to take the leap. They must embrace AI.