How Community Banks Can Modernize Their Tech Stack

When Anthropic’s Claude Mythos, a security-focused AI model, previewed this spring, many community bankers were shocked by how AI could soon target decades-old vulnerabilities in their existing code based on how that code was written. 

The arrival of Mythos and other groundbreaking AI solutions is a powerful example of a seismic change that is spurring community banks to reassess their technology stacks. 

“I could suddenly see us investing in technologies that weren’t in the budget, because when we came up with our annual budget in late 2025, these threats didn’t exist,” says Anthony Ranghelli, chief information officer at $975 million-asset Potomac Bank in Charles Town, West Virginia. In fact, management consulting firm Bain & Company says the average company should increase its cybersecurity spending by two or more times their current levels given what AI is unleashing.

Ranghelli sees the tech space evolving so rapidly that “a true technologist at a bank is probably assessing tech needs every day.”

Should tech drive bank strategy?

Community bankers are finding that assessing their bank’s tech stack no longer fits neatly into the traditional budgeting paradigm.

One of the biggest changes, according to Wayne Miller, ICBA’s chief innovation officer, is that now more than ever, “the tech stack should reflect a bank’s strategy.” Understanding this, bankers are resisting shiny new objects, insisting that tech investments alleviate stubborn problems and help the organization reach its strategic goals.

Andrew E. Silsby, president and CEO of $1.8 billion-asset Kennebec Savings Bank in Augusta, Maine, exemplifies this trend. He relishes the fact that business line managers are now bringing tech ideas to the table. He describes the tech spend at his bank as both more dispersed and more strategic. 

“Nowadays, you don’t start with the technology out there,” Silsby says. “You have to start with the problem you’re trying to solve.”  

He notes that allowing bank employees to propose tech projects generates a broader range of ideas. “I want the residential lending department to say, ‘Hey, we’ve got a potential new solution here. We think we can save some money and be more efficient,’” he says. “When that happens, it’s beautiful.”

However, community banks are often constrained by outside forces, such as the customary timetables for partnership agreements. 

While most fintechs strive to lock in one- to three-year contracts, the average core banking contract runs for five to seven years, says Christian Ruppe, partner and chief strategy officer at tech consultancy impactFI Advisors in Atlanta.

Miller notes that even when a core contract is nearing its renewal date, many community bankers balk at making a change because of a widely acknowledged truth: Implementing a new technology can be daunting. He says that unlike other investments within the tech stack, core conversions have been likened to “a skeleton transplant” because of the sheer pain and disruption involved.

New evaluation processes

As the landscape evolves, some tech projects appear more appealing and others less so. 

Gone are the days when assessing a tech stack meant annually assembling key employees who “huddled around a table and got the job done,” says Silsby.

Today, any departmental Kennebec Savings Bank employees can propose a new technology expenditure through a process that starts with an intake form and a project plan. For all projects of considerable scale, a member of the leadership team must agree to act as a sponsor. Silsby says someone who’s “good at assigning tasks and nagging people in a nice way” must sign on too.

Another telling change at Kennebec is that prospective projects are now accompanied by a proposed budget, including implementation costs. “Notifying the IT and finance departments about major projects coming down the pike means that nobody is surprised anymore,” he says.

Silsby notes that even slotting approved technologies into the calendar has become far more strategic, with details such as when other tech projects go live and when key players are taking vacation now part of the planning calculus.

Cal Roberson, vice president of the financial institution division at IT managed-services provider Integris in Cranbury, New Jersey, maintains that scheduling can make or break a project. He says February through early June and late August until October are choice seasons for tech implementations, because employees tend to be in the office and can tackle larger lifts.

Roberson says another key aspect of scheduling is the impact on customers. Going live around major holidays can work well when a tech implementation is extremely disruptive, because fewer users are transacting business.

Through his experience as former chief innovation officer at $3.7 billion-asset Colony Bank in Fitzgerald, Georgia, Ruppe learned the value of upgrading to a third-party digital banking platform, especially for institutions with ambitious growth strategies and a healthy sales culture. He’s seen that “changing to a more modern [digital-banking] platform is a terrific way to increase assets.”

On the other hand, Ruppe recommends reassessing investments in data visualization tools. He says that given the rise in AI, banks increasingly have the wherewithal to build their own KPI dashboards and analyze data without outside help.

A shift toward AI

Discussions about AI have quickly started dominating tech assessments.

Silsby notes his bank is leaning heavily into AI, beginning with a bank-wide move to incorporate Microsoft Copilot into daily tasks.

For example, Silsby now uses Copilot to plan his workday. After linking to his calendar and inputting his home address and the addresses of the bank’s operation center and branches, he asks Copilot to organize his day for maximum efficiency. “If I’m going to encourage employees to lean into AI, I’m going to lean in, too,” he says. 

What’s more, to show management the breadth of new possibilities, Kennebec Savings Bank recently purchased six laptops that operate independently from the bank’s network, as well as subscriptions to eight of the leading AI platforms. “We’ve passed out laptops to our leadership and said, ‘Go play,’” says Silsby. As one example of what AI can enable, he has started creating a tool so he and his commercial lenders can measure the profitability of any loan based on various interest rate pricing options. Here, the goal is for lenders to learn which loans will make the bank money—and which won’t.

Silsby is hardly alone in believing that a spirit of experimentation is critical to incorporating AI into a bank’s tech conversations.

“If you learn how to use AI, you’ll quickly change over to new models,” says Ruppe. “It’s way more important to build the muscle of using AI than to determine how you’ll be using AI three years from now.”