Banks have long touted generative AI's possible benefits for client service or internal operations, but financial institutions are now considering how the technology might help free up time for CEOs and their teams.
Leaders of small financial institutions, particularly credit unions and community banks, are hoping generative AI will help them punch above their weight.
"I get hundreds of emails a day, even something as simple as relying on [Microsoft] Copilot to prioritize and organize my inbox right away, that's a half hour to an hour out of the day where I can be doing more high-value things, either supporting my co-workers, or just engaging more with people outside the bank," said Chuck Leach, president and CEO of Lee Bank, a community bank in Lee, Massachusetts.
The bank, which has around 85 employees, plans to roll out generative AI more broadly across its employee base in late 2025 or early 2026, he said.
In the interim, a small group of employees, including the CEO and IT professionals, are testing the use of ChatGPT to explore how generative AI might improve their work outcomes. For now, they're keeping expectations in check, with use cases that include helping team members review and analyze emails; researching regulations; and offering a first draft of communications or correspondence.
"If we're looking to, you know, figure out how to brand something, or finding out how maybe a regulation has changed … [or] crafting a paragraph if it's more going out to a broader audience," he said.
A longer road to adoption among leaders
While generative AI adoption is growing across banks — an October
Christine Livingston, managing director and leader of AI and Internet of Things practices at consulting firm Protiviti, identified two key areas where AI could impact banks and bank CEOs. They include decision support — synthesizing and contextualizing information — and the quicker retrieval of information and data to facilitate faster decision making.
"I'm seeing some really interesting progress on immediate decision making and ability to do scenario simulation pretty quickly," she said. "Rather than having that point-in-time analysis … what if I have this continuous stream of insights? We can ask many different scenarios that can be simulated and modeled almost instantaneously."
It may take around five years to take data-driven decision making capabilities to scale, however, with companies typically taking small steps by introducing conversational AI and analytics capabilities, she noted.
Over time, Leach said he envisions an environment where generative AI helps the CEO access data to identify new client acquisition and retention opportunities.
"If we can leverage the data, instead of doing it the old-fashioned way and going through the files and trying to put it in a spreadsheet, [if we can] figure out a way to use generative AI to identify those opportunities, that would be tremendous for us," he said.
AI and the office of the CEO
Michigan State Federal Credit Union is currently exploring wider adoption of generative AI across business units, including the CEO and staff members who support her. Since June, the credit union has been allowing a pilot test group of 100 staff members to explore how generative AI through Microsoft Copilot can deliver efficiencies for the organization, said Benjamin Maxim, chief innovation officer.
So far, most of the use cases aren't revolutionary; they include drafting emails and corporate documents and generating code. But Maxim said AI could transform the work of the CEO and employees who support her.
Take, for example, the work of the vice president of executive communication. This executive supports senior leadership by drafting blog posts, corporate messages and other written communication content on behalf of the CEO. With generative AI preparing a first draft, it could cut down considerably on preparation time, allowing the CEO and the vice president of executive communication to spend more time with stakeholders.
Generative AI "allows for more time to do the relationship building, because I ultimately think that one of the core jobs of the CEO is relationship building and board management," said Maxim.
While he doesn't see the communications executive role being replaced by AI, it will allow that person to focus more on relationship building in the community, he added.
Support staff
Staff that report directly to the CEO may see their work change rapidly as AI systems become more widely adopted. Leach, who has six direct reports, said using AI tools to prepare meeting presentations for board meetings alone could save around 100 hours of employee time per year.
"We have a monthly board meeting, and we have other PowerPoint driven meetings, and I would say hours go into straightening out the PowerPoint each month and standardizing it," said Leach. "Copilot will really help drive efficiency there," cleaning up documents and standardizing them far more quickly than humans can.
Asked whether AI will reduce the number of staff members that report to him, Leach said while AI may change their jobs, it won't reduce the team footprint.
Managing AI systems CEOs use is likely to be a new responsibility for teams that report to CEOs.
"You might, in the future, have someone on staff who's responsible for managing the CEO's AI, who is thinking about how we make sure it's connected and contextual and who is monitoring and tracking the AI's performance," Livingston said, noting that an AI assistant's work may need to be monitored the same way human worker deliverables are evaluated.
Guardrails and 'human in the loop'
As generative AI becomes a bigger part of the role of a CEO and across the organization, institutions say they're building governance and review frameworks to minimize data leakage and other cyber challenges, along with mistakes that come from hallucinated input.
Michigan State Federal Credit Union set up an AI governance committee last year with representatives across the institution that reviews AI use cases and ensures guardrails are in place.
"Any new generative AI technology needs to be reviewed as part of their due diligence process with this committee to ensure that guardrails are in place, that we're not leaking sensitive or proprietary information into a system," Maxim said.
Any employee whose byline appears in a communication prepared with AI assistance needs to review it before it can be finalized. In some cases, the committee will need to review text developed with AI assistance, he adds.
In addition, if CEOs and their teams are using AI more frequently, institutions need to build regular review mechanisms — AI audits — to verify model accuracy, rate of hallucinations and bias and other benchmarks. Michigan State Federal Credit Union is in the process of developing a third-party review process for AI systems that would encompass AI carried out by the CEO and their team.
"There needs to be integrity checks. There are consulting services that are starting to spring up to help you with some of that integrity and model validation," Maxim said, with data leakage and cyber vulnerabilities evaluated as part of that assessment.
