Shanthi Shanmugam is the co-founder and CEO of Casap, an agentic AI platform that simplifies dispute management for banks and credit unions.gettyPicture a fraudster. You might imagine someone in a hoodie, hunched over a laptop, running code from a foreign server. However, that image is outdated, and the reality is that a lot of fraud today is committed by legitimate, everyday customers. It’s a phenomenon called first-party fraud, and it's costing the financial industry $100 billion a year.​According to LexisNexis Risk Solutions' Cybercrime Report, first-party fraud now accounts for 36% of all reported fraud, up from just 15% the year prior. And, according to the survey cited below, nearly half of financial institutions saw fraud losses increase in 2025, with two-thirds expecting them to rise again this year.To better understand how financial institutions are responding, Casap partnered with Cornerstone Advisors to survey more than 400 bank and credit union executives, and what we found was a system buckling under pressure in three key ways: institutions can't detect the fraud that's already happening, fraudsters face almost no consequences for trying and the tools being deployed weren't built for this sort of threat.​The system was built to trust customers​.There is a catch-22 at the heart of first-party fraud. The dispute system was designed to protect consumers, but when the customer is the threat, that protection becomes a mechanism for loss.​Most financial institutions issue provisional credit before completing a meaningful investigation. That's not a policy failure, but rather an effort to adhere to strict Reg E compliance timelines. That problem is compounded by the fact that most institutions don't have tools capable of distinguishing a legitimate claim from a fraudulent one before credit is issued. With this in mind, it’s not surprising to hear that only 9% of bank leaders surveyed are “very confident” their organization captures all fraudulent activity. In fact, nearly half don't know, or don't track, what percentage of false claims are caught before provisional credit goes out the door.​Once that credit is issued, it rarely comes back. The investigation clock is running, the team is overwhelmed and the loss has already been absorbed. That's the math that serial dispute filers are taking advantage of.​There is no reason to stop​.This isn’t just about individual bad actors. The problem is that the incentive structure is broken. A whopping 91% of bankers believe low perceived consequences embolden false filers. And they're right because the system validates that belief every time it issues credit without asking questions.​But it's not just opportunism. The data points to a convergence of factors making this problem stickier than any single rule change can fix.​Nearly 90% of bank leaders believe financial hardship is driving suspicious disputes. People in financial distress don't always think of disputing a legitimate charge as fraud, but as a means for survival. And then there's the cultural layer: Around 60% of bankers say social media content has made customers more comfortable filing questionable disputes. I watched this play out in real time when the Chase ATM “glitch” went viral on TikTok in 2024. Deposit manipulation became a tutorial, and the institution's own infrastructure became the attack surface.​Institutions that frame this as a purely criminal problem will continue underestimating it.​Utilize smarter detection without sacrificing customer experience.​The fraud detection stack most institutions rely on was built to stop third-party fraud, which focuses on stolen credentials, account takeovers and synthetic identities. But because the login is legitimate and the customer is who they say they are in first-party fraud cases, there's no anomaly to flag at authorization. The fraud shows up later, buried in a caseload that's growing faster than teams can manage.​Our data shows the top operational challenge for teams is distinguishing legitimate claims from fraudulent ones. The use of manual processes, lack of system integration and case volume intensify the problem. The infrastructure was simply built for a different era.​The answer isn't to treat every customer like a suspect. The challenge—and the opportunity—is adding accountability without punishing the honest majority. Most customers filing disputes are filing them legitimately. The goal is to make fraudulent disputing harder for the dishonest few, without degrading the experience of the ethical many.​There are tools purpose-built to detect first-party fraud. AI-native platforms that analyze behavioral patterns, dispute history and transactional context can surface signals that transaction-level monitoring was not designed to catch. Still, hesitancy remains, as 42% of bank leaders say their organizations are uncomfortable deploying AI for evidence gathering, decision support or customer communication. Despite this uncertainty, 70% see potential for agentic AI in fraud detection.Accountability is a customer experience strategy​.The face of fraud has changed. The response must, too.​Financial institutions that treat first-party fraud as purely a criminal matter will keep losing ground. The future belongs to those who recognize it as a behavioral, systemic, technological issue that needs to be solved through the customer experience lens. Institutions that pair advanced detection tools with clear communication about consequences and financial wellness resources will be better positioned to reduce recidivism without losing customers.​Building trust in the next era of banking isn't just about faster payments or better apps. It's about making sure the system is fair for the customers who rely on it and accountable for those who exploit it. ​Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?