AI fraud fears rise as payment scams hit finance teams

Yooz has published research showing that payment fraud attempts are now a common risk for finance teams. In the survey, 70% of U.S. finance professionals said their organisation had experienced a payment fraud attempt in the past two years or could not rule one out.

The findings are based on a survey of 750 U.S. professionals working in finance, accounting, or accounts payable roles. The study also points to growing concern over AI-driven fraud, with 48% ranking threats such as deepfakes and impersonation among their biggest emerging worries.

Trust appears to be a weak point in payment approval processes. Nearly half of respondents, 49%, said they had approved or almost approved an unusual payment request without extra verification because it appeared to come from a trusted source.

This suggests social engineering remains effective even as finance teams become more aware of digital threats. Pressure to act on requests from senior colleagues, suppliers, or other familiar contacts can override normal checks.

Losses from known fraud attempts were also significant. Among organisations that had experienced a known fraud attempt, 28% lost money. Of those incidents, 39% led to losses of USD $50,000 or more.

The results also show a link between manual processes and fraud exposure. Finance teams with mostly manual accounts payable workflows lost money in 42% of known fraud attempts, compared with 30% for highly automated teams and 22% for those using a mix of automated and manual processes.

AI concerns

Respondents said AI is making fraud harder to detect. While 48% identified AI-powered threats as their top emerging concern, only 21% said finance teams were extremely prepared to deal with AI-enabled fraud.

The gap between concern and preparedness is notable because the same research found that teams already using AI were more likely to identify attempted fraud. Finance teams actively using AI detected fraud attempts at more than twice the rate of non-users, 63% compared with 30%.

The report presents that contrast as a challenge for finance departments that still rely heavily on manual review and staff judgement. As fraudulent emails, invoices, and voice messages become more convincing, traditional checks may be less effective unless backed by stronger controls.

Payment fraud has become a broader operational issue rather than a narrow compliance problem. The findings suggest weak points can emerge across invoice handling, payment approvals, and supplier communications, especially when requests appear routine or come from recognised contacts.

For accounts payable teams, the research adds to wider concerns over impersonation fraud and business email compromise. These schemes often depend less on technical intrusion and more on persuading staff to bypass verification steps.

The study examined how organisations experience, detect, and respond to payment fraud, as well as the role of automation and AI in prevention. It also looked at where finance teams remain most vulnerable as fraudulent requests become harder to distinguish from legitimate ones.

Laurent Charpentier, Chief Executive Officer of Yooz, said the problem has moved beyond basic employee awareness training.

"For years, companies have treated fraud mainly as a people problem: train employees to spot the fake email and hope an approver catches something suspicious," Charpentier said.

He said the rise of generative AI is changing that balance. "That approach is becoming less reliable as generative AI makes fraudulent invoices, emails, and even voice requests much harder to distinguish from the real thing. Fraud attempts will continue to become more sophisticated, but stronger processes can prevent more fraudulent payments from going through," Charpentier said.