Financial institutions are deploying AI to counter internal fraud attempts as fraudsters look to bank employees to do their dirty work. 

“Insider recruiting” is becoming a major problem for FIs, Lenny Gusel, North American head of fraud solutions at fraud mitigation service provider Feedzai, told FinAi News. 

“In simple terms, it is fraud syndicates and transnational crime rings recruiting frontline bank employees to commit fraud on their behalf,” he said. 

The fraud usually is very simple, Gusel said, adding that most of the time, the fraudster would ask employees to steal personal identifiable information of a customer, remove a freeze on their account or checks, or send password reset emails or messages, which are intercepted by fraudsters. 

While the problem is not new, it has been gaining traction in recent years as employees struggle financially and have an incentive to make something extra on the side, Gusel said. 

According to ACFE’s Occupational Fraud 2026 report published in May, the median internal fraud loss to banks between September 2025 and April 2026 was $100,000, with the average being $1.5 million.  

Almost half of the perpetrators had worked for the victim organization for one to five years, the report stated. Perpetrators with a tenure of at least 10 years caused the highest median loss. 

According to a May U.S. Department of Justice release, one bank insider pleaded guilty to facilitating fraud schemes for outside conspirators defrauding TD Bank. 

Cheungkin Lam “abused his position as a bank employee to help fraudsters steal money from unwitting customers and bribed another bank employee to do the same,” Assistant Attorney General A. Tysen Duva said in the release. 

Gusel said Wells Fargo and Regions Bank have been victims of similar fraud. 

AI can help 

Machine learning has been in the fraud fighting toolbox for decades, Gusel said, adding that banks have deployed it externally and internally to spot unusual behavior. 

A traditional machine learning model will keep an eye on what employee is opening what files in a bank, and, if there is anything suspicious, it will alert authorities, he said. 

“It’s not the only way, but that’s the kind of AI that has the most prevalence here,” Gusel said.  

Feedzai, which works with over 100 FIs globally, is evaluating deploying large tabular models (LTMs) to keep such events in check, Gusel said. 

What are LTMs? 

LTMs are gen AI models that are trained to process and analyze structured data in rows and columns, according to AI infrastructure platform Granica AI. 

That specific LTM training can help AI better predict outcomes, Gusel said. 

“That data can be financial data, someone’s click rate when going through a website or anything can be portrayed in a tabular format,” he said. 

Feedzai is in the early stages of developing LTMs, Gusel said, adding that many FIs are interested in such offerings. 

Layers of intelligence 

AI can be used to run constant background checks on employees to see if they are collaborating with fraudsters, Pierre Buhler, managing director of financial services practice at consultancy SSA, told FinAi News. 

Bank security teams can also use AI to monitor click rates for suspicious patterns, rather than being reactive and blocking files after they have been accessed, he said. 

FinAi Lending Summit, set for Oct. 7-8 in Las Vegas, will include speakers from Fifth Third and Capital One as well as a fireside chat with Piermont Bank founder and Chief Executive Wendy Cai-Lee. To learn more about the 2026 event and register for early-bird pricing through Sept. 4, visit here.