Fraud disputes are one of the fastest ways for banks to lose customers—and one of the least prioritized parts of the business. Despite the high costs, many institutions still treat them as a back-office function rather than a decisive point in the customer relationship.

Beyond immediate losses—such as chargebacks, write-offs, and investigation expenses—banks also lose revenue when an engaged customer no longer keeps their account top of wallet.

In a PaymentsJournal Podcast, Steve Durney, Vice President of Partnerships and Alliances at Quavo, and Suzanne Sando, Fraud Analyst at Javelin Strategy & Research, discussed the hidden costs of fraud disputes. It’s a problem that will only intensify as AI and agentic commerce evolve.

Customers Are Willing to Move On

Banking industry recovery rates for fraud average around 64%, leaving more than a third of disputed dollars unrecovered. For most banks, disputes are an expensive process, with operating costs eroding already thin margins.

Research suggests that if a fraud issue or dispute isn’t handled effectively, 60% to 70% of customers will move to another bank. Notably, the outcome doesn’t always have to favor the customer, as long as the process is managed transparently and resolved efficiently.

Once customers feel they aren’t being treated fairly, it’s difficult to restore trust. Accounts may be quietly abandoned, and products go unused. Even without formally closing an account, customers often disengage entirely.

“We’re seeing growing numbers of consumers who are willing to close an account and walk away when they have a bad experience with their account,” said Sando. “Setting everything back up with a whole new financial institution—like bill pay or getting all your accounts linked to whatever other financial accounts you were linked to—it’s a tremendous hassle. If you’re willing to go through all of that, that says a lot for how important security and customer service is throughout a process like this.”

Modernizing the Dispute Process

Several aspects of the dispute process need modernization to improve efficiency and recover lost value. Because dispute teams rarely receive priority budget allocation, banks often underinvest in technologies that could significantly improve performance.

Organizations that ignore these inefficiencies and continue to deprioritize back-office enhancements only prolong the problem. Five years from now, they are likely to be facing the same challenges.

“The inefficiency really comes into what historically would have been categorized as judgement—something where a human being has to give opinion on in order to route it properly,” said Durney. “Second is the document interpretation, for documents that are incoming from either a consumer or a merchant, or being transmitted from the bank to the merchant. That’s the lion’s share of the inefficiency.”

Banks without standardized documentation and clear rules force teams to spend valuable time interpreting procedures instead of executing them. When deeper, manual investigations are required, staff should be freed from repetitive administrative tasks so they can focus on higher-value work.

Fighting Against Constant Turnover

High turnover within fraud teams is another persistent challenge, especially given the long ramp-up time required for investigators to become effective.

“I asked a bank not long ago, ‘What’s the turnover rate in your department and how long does it take you to onboard somebody?’” said Durney. “They said the onboarding was about six to seven months before they were effective, and they had a turnover rate of roughly 25%.

“If you’re turning over your staff every four years and it takes you six to nine months to have somebody be a top performer, that’s a radical impact on all these day-to-day manual tasks,” he said. “There has to be a way to get people up to speed faster, handling the cases in such a way that you can actually hold on to staff and you don’t have that turnover.”

Involving experienced compliance and regulatory professionals in designing dispute process technology can help reduce risk and ensure systems are better equipped to handle complex scenarios.

The Risks of Agentic Commerce

While banks are still working to modernize dispute processes and stabilize fraud teams, the next wave of change is already emerging.

Agentic commerce promises new opportunities, but also introduces significant fraud risks. When AI agents act on behalf of consumers, traditional fraud signals—such as behavioral biometrics, device intelligence, and IP address—become less reliable, making it harder to distinguish legitimate activity.

Fraudsters will increasingly leverage agentic AI in ways that are difficult to predict.

“Once people really figure out how to use the tools to be able to make the agents go off and do things, you’re going to get the gray area of people abusing the system,” said Durney. “The use cases that we see so far are using AI to navigate the system. Say: ‘I bank with Bank X, tell me how to navigate the disputes process,’ and it will generally give you a pretty good recipe as to how to get through it.”

Banks are already anticipating how AI could go wrong. Those looking to stay ahead in fraud dispute management must prepare now.

“HBO’s Silicon Valley has a perfect example of this,” Durney said. “They told the AI to go buy them burgers for lunch, and then a pallet of frozen hamburgers showed up. Did the AI do what it was supposed to do? More importantly, what is a consumer going to do? A consumer is going to find the easiest path to go. I need somebody to be my advocate to fix this problem because I didn’t want a pallet of frozen hamburgers.”