Much like the introduction of the internet in the mid-1990s and the smartphone revolution 15 years ago, AI has the power to fundamentally redefine how customers engage with their financial institutions. This was a recurring conviction at last month’s NEXT Forum in Philadelphia, where a panel of three forward-thinking leaders helped an audience of bank executives unpack this sprawling subject.
LISTEN HERE: YOUTUBE | SPOTIFY | APPLE PODCASTS
1. Start Small, but Start Now
From the stage, Bankwell’s Chief Innovation Officer, Ryan Hildebrand, encouraged community banks to begin exploring AI with low-risk, high-reward pilots. A seminal example of this among banks today is the practice of rolling out Microsoft Copilot or similar tools in controlled environments. Hildebrand described a practical approach, starting with a small pilot team that includes representatives from finance, marketing, operations, and risk. “We identified people that were really excited about using AI,” he said, noting that the group was focused on identifying ways their teams could incorporate the technology day-to-day. The AI team met weekly to share use cases and expanded them based on results.
While the pilot team was a great catalyst, Hildebrand noted that AI preparation has to start well before a bank launches use cases. “It starts with the data,” he explained. “If you haven’t invested in data, lean into that. It’s really important [because], if you can’t track something, then you’re not going to be able to figure out what that ROI is.”
ROI tracking is crucial for AI efforts, Hildebrand says. “One of the big issues that we see is not understanding what value you’re actually getting. It may be cool, but if you’re not getting value from it, then you’re wasting time as well as money.” For every AI project at Bankwell, leadership creates a business plan that includes projected ROI — a measure that banks need a strong data foundation to be able to track.
Takeaway: Success with AI isn’t about massive investments up front; it’s about cultivating internal champions, testing use cases, and building a culture of curiosity and experimentation.
2. Make Sure Your AI Efforts are Explainable
As banks are rolling out new projects, it’s also important that they be able to document and explain the bank’s AI use. Jeremy Mandell, Co-Chair of the Financial Services Practice at Morrison Foerster, noted that this gets to the core of ‘explainable AI.’ Mandell shared that, “a lot of times when a regulator or an exam team is trying to learn how banks are deploying AI, they are learning along with their exam process.” That’s why it’s critical to be able to communicate the details in a clear way.
Mandell described explainable AI as “the ability to say the purpose of the AI, understanding the inputs that are used, the outputs and the suitability of those outputs, and the business justification of the use of that AI.” He encouraged Forum attendees to articulate all of those things and document them in a way that is not only understandable for management, but also for regulators and exam teams.
Takeaway: Banks need structured documentation and a firm handle on the inputs and outputs of their models for AI deployments to ensure transparency for regulators and accountability for the board.
3. AI Should Enhance, Not Replace, Human Decision-Making
With much of the emphasis in this area on efficiency, it’s interesting that it was the technologist of the group, Chief Revenue Officer at Alkami Technology & Co-Founder and President of MANTL Nathaniel Harley, who highlighted AI’s role in deepening relationships with customers.
Harley painted a picture for the audience of part of his AI vision, saying “as we think about deepening relationships, it can really make the banker more intelligent, right? You have someone sitting in front of you. You have all this data [about them] that might be in different systems. How do you aggregate that data and then serve the banker information so that they know the next best product that they should offer to their customer or business that’s sitting in front of them?”
While noting that AI can provide efficiency gains in areas like fraud detection and manual reviews, Harley also stressed the importance of keeping a “human in the loop” to validate AI outcomes.
Takeaway: AI’s true value in banking lies in augmenting — not eliminating — human expertise, improving productivity, and enabling smarter, faster customer engagement.
