ProSight survey says fraud is now a human problem too

ProSight Financial Association published its 2026 State of Fraud Prevention survey, which found that financial institutions increasingly see fraud as both a human and a technical problem.

The survey of more than 125 executives at North American financial institutions found that gaps in customer and employee fraud awareness were viewed as the biggest sources of preventable losses. Some 62% of respondents pointed to customer awareness gaps, while 49% cited employee awareness gaps.

More than half of respondents said fraud prevention budgets had risen by at least 5% in the most recent fiscal year. Even so, 85% said the threat environment was changing faster than their organisations could respond.

The findings suggest that the main pressures on anti-fraud programmes are operational rather than purely financial. Alert overload was cited by 64% of respondents and data quality by 62%, compared with 51% who identified budget constraints.

Planned spending over the next three years still leans heavily toward technology. Identity verification led expected increases at 46%, followed by real-time payment controls at 43% and artificial intelligence and machine learning tools at 41%.

By contrast, internal staffing ranked last among eight areas for planned increases, with only 8% of respondents expecting to raise investment there. This points to a gap between rising fraud-monitoring demands and the number of people available to review cases.

Threat shift

Check fraud emerged in the survey as both the largest source of loss exposure and the fastest-growing fraud type. It was cited by 58% of respondents as the biggest source of exposure and by 46% as the fastest-growing type of fraud.

Authorised push payment scams ranked second for growth, with 40% of respondents identifying them as one of the fastest-rising threats. These scams are especially difficult for institutions because they rely on persuading customers to move money themselves.

Jason Bartolacci, Director of the ProSight Fraud Alert Network, said criminals had adapted as banks strengthened their digital defences.

"Banking institutions have spent years hardening their digital channels, and criminals have adjusted. They are going after the customer directly through bank and law enforcement impersonation scams. When a customer authorizes the payment themselves, a detection model has far less to work with. You are asking technology to catch a decision that a person has already made, which makes the solution increasingly difficult for the institution," Bartolacci said.

The survey also highlighted the time pressure created by faster payment systems. Once a real-time payment has been sent, recovery becomes much harder, increasing the importance of intervention before funds leave an account.

Collaboration gap

Interest in sharing information with peer institutions appears high, but actual practice remains limited. Nearly two-thirds of respondents, or 64%, said they were very interested in deeper information sharing with peers.

After a notable fraud incident, however, only 24% said they shared information directly with peer institutions, and 20% said they did so through formal industry networks. By comparison, 71% said they report such incidents to law enforcement.

Jen Guidi, Chief Executive Officer of ProSight Financial Association, said the underlying problems were difficult for institutions to solve alone.

"These are problems institutions are unlikely to solve one at a time, which is part of why we've seen such a strong appetite for working with peers. Our role is to be the trusted place where that occurs," Guidi said.

Customer focus

The results indicate that many institutions are putting more emphasis on stopping fraud before customers act. Some 87% of respondents said they plan to increase investment in customer education, while 77% said they already use direct customer communications as part of fraud prevention efforts.

This reflects a broader shift in fraud prevention strategy, as institutions try to address scams that begin outside bank systems and depend on social engineering rather than technical intrusion. It also underscores the difficulty of relying solely on back-end controls once a customer has been persuaded to authorise a transaction.

Bartolacci said the wider fraud picture is often fragmented across institutions.

"Fraud is a network problem. No single institution sees the whole picture, and the schemes that do the most damage usually hit several institutions before anyone connects them. Fraud teams understand that. The ProSight Fraud Alert Network was built to help the industry combat these threats together by creating a vetted place where fraud professionals can learn about these fast-evolving threats from each other before those threats spread," he said.