Financial institutions and fintechs are recognizing AI’s ability to improve overall portfolio health, whether it’s through reduced delinquencies, enhanced credit decisions or diversification. 

In an FDIC case study, AI decreased loan default rates by 2.7 percentage points, down nearly 30%, according to a 2025 report by the agency, which evaluated a dataset of more than 4.5 million loans.  

Slashing delinquencies  

AI’s ability to detect patterns and predict late payments can strengthen portfolio health, Michael Orsomarso, senior vice president of data and analytics at fintech Akuvo, told FinAi News. The company provides credit risk and debt collection solutions. 

“Instead of being reactive, you can be proactive and say, ‘If I know that this is projected to be a delinquency that’s going to span multiple payments, maybe I can do a hardship solicitation up front, maybe I can work with them up front,’” he said.  

The fintech offers a solution for managing early-stage delinquencies, using AI to analyze data, identify risk patterns and automate outreach to borrowers, enabling FIs to “nip it in the bud from Day 1,” Orsomarso said.  

Akuvo’s model is trained on account history to help lenders predict whether a delinquency will self-cure or escalate, Orsomarso said.  

Customer-facing agentic AI tools also reduce delinquencies, in part because they can operate 24/7, Benjamin Maxim, chief technology officer at Michigan State University Federal Credit Union, previously told FinAi News. 

In April, the $8.3 billion credit union deployed fintech Constant AI’s Skip-a-Pay agent.  

“Say you’re panicked at night like, ‘I need to pay this loan that’s due tomorrow.’ You now have a real option and someone to talk to,” Maxim said. 

Portfolio diversification 

AI also delivers value in portfolio management and diversification, Will Rhoads, chief innovation officer at Sonata Bank, told FinAi News.  

The $300 million bank finds AI especially beneficial in its loan portfolio for quick-service restaurant franchisees, Rhoads said. 

“How a given brand is performing across the national footprint is a signal that we want as a franchise lender, but we rarely get the specificity that we can build on. With [AI], we’re looking to take something that could become a risk and turn it into a specialty, where we really understand things at a more micro level.” 

— Will Rhoads, CIO, Sonata Bank

Rhoads said AI is effective for evaluating portfolio metrics including:

• Cash-flow patterns;
• Transaction velocity;
• Cost accounts; and
• Influences of broader market conditions.  

“AI lets us build insights and scale around that without necessarily having to have all these domain experts,” he said.  

Simulating risk 

AI also can enhance credit and portfolio decisions by providing deeper cash-flow analysis and simulating real-word scenarios, David Snitkof, general manager of SMB at lending fintech Ocrolus, told FinAi News.  

“A good risk manager looks at many different outcomes that could happen,” he said. “Let’s simulate all of those different versions of the future and see if they would cause me to take a loss. … Then you try to see what you would do to mitigate that so you’re not at risk.” 

AI agents have the potential to operate as “talented risk analysts, constantly looking at every corner of your portfolio and asking those questions, running those scenarios, stress testing,” Snitkof said.  

Register here for the FinAi Lending Summit, set for Oct. 7-8 in Las Vegas.