EMEA fraud attempts rise as AI scams grow more complex

FICO has published research showing that fraud attempts are rising and becoming more complex for financial institutions in Europe, the Middle East and Africa. More than a third of respondents reported an increase of more than 25% in fraud attempts over the past 24 months.

The study surveyed 202 senior fraud, risk and technology professionals at retail banks, neobanks and fintechs. It found that 52% of EMEA institutions now see growing fraud complexity as their main fraud prevention challenge, compared with 40% globally.

Almost half of respondents in the region also identified effective AI integration as a leading challenge. In EMEA, the figure stood at 48%, broadly in line with the global response, suggesting institutions are still struggling to fit new AI systems into existing fraud controls.

Losses are also climbing for a sizeable minority of firms. Some 28% of EMEA institutions said financial losses from fraud had increased by more than 25% over the past two years.

Threat picture

The findings highlight particular concern in the region over AI-enhanced fraud and social engineering scams. EMEA respondents rated AI-enhanced fraud at 4.42 out of 5 and social engineering scams at 4.36, both above the global averages of 3.86 and 3.90.

Synthetic identities and deepfake fraud also ranked highly among regional concerns. Those threats scored 4.02 and 3.96 respectively, again above global levels of concern.

Adam Davies, Vice President, Product Management, FICO, said the issue is less about raw growth in fraud than the changing nature of attacks. "Fraud in EMEA is not accelerating out of control, but it is becoming far more sophisticated, and that shift matters just as much," Davies said.

"Fraudsters now have access to the same generative and agentic AI tools that banks are deploying to stop them, and they are using them to scale social engineering, synthetic identities and deepfake-enabled attacks. The institutions pulling ahead are the ones connecting their fraud functions across the enterprise and orchestrating every intelligence source they have, rather than adding point solutions one at a time," he added.

Strategy shift

The survey suggests many institutions are responding by taking a broader approach to fraud management. Some 86% of EMEA respondents said a unified, enterprise-wide fraud strategy is essential, with 38% treating it as a critical priority and 48% pursuing it as a longer-term focus.

That shift comes as firms try to manage the trade-off between tighter controls and customer friction. Half of the EMEA institutions surveyed reported false-positive rates above 15%, indicating that large numbers of legitimate customer actions are still being flagged.

The research also points to growing dependence on outside suppliers. Third-party vendor models now have the greatest influence over fraud decisions at 46% of EMEA institutions, compared with 24% for in-house models.

Davies said the next phase will depend on how well firms connect multiple sources of intelligence. "The next 24 months will be defined by orchestration, not by any single piece of technology," he said.

"EMEA institutions have already built strong foundations in AI-driven detection, and confidence in agentic AI is particularly high in the region. What's needed now is the connective tissue: bringing in-house models, vendor intelligence and consortium data together so fraud teams can act on a complete picture in real time, without pushing more friction onto legitimate customers," Davies said.

FICO said financial institutions should focus on moving AI systems from pilot projects into wider operational use while linking fraud controls across models, channels, products and portfolios. It also argued for risk-based investment decisions guided by financial impact and regulatory exposure.

The research indicated that firms are increasingly viewing fraud prevention and customer experience as linked issues rather than separate objectives. "You should prioritize what works within disciplined fraud control frameworks rather than launching yet more pilots," Davies said.

He added that institutions that can reduce fraud while limiting disruption for legitimate users will be in a stronger position. "The organizations that can do both will be the ones that define competitive advantage in this era," he said.