AI now default tool in fraud & AML, says SEON report

SEON has published new survey findings suggesting artificial intelligence is now a default tool for fraud prevention and anti-money laundering (AML), even as staffing and budgets continue to rise across the sector.

The AI Reality Check: 2026 Fraud & AML Leaders Report draws on responses from 1,010 fraud, risk and compliance leaders at director level or above. Respondents work across payments, fintech, financial services, retail, eCommerce and gaming, spanning North America, EMEA, Latin America and APAC.

The results point to near-universal AI use in day-to-day workflows. In the survey, 98% of organisations said they integrate AI into daily fraud and AML work, with only 2% still in the planning stage. Another 95% said they are confident AI can detect and prevent fraud, including 52% who are very confident.

At the same time, respondents reported continued investment in people and systems. The survey found 94% plan to add at least one full-time fraud or AML hire, up from 88% a year earlier. Another 83% expect fraud and AML budgets to increase during 2026.

Pressure on losses

One of the clearest year-on-year shifts was how leaders view the gap between fraud losses and revenue growth. The number who disagreed with the statement "fraud losses are growing faster than revenue" fell by almost 40% from the previous year. That change suggests growing concern that losses are closing in on growth rates for many organisations.

SEON framed the findings as evidence that AI adoption alone is not making operations easier. Threat volumes and complexity are still increasing, it said, and criminals are also using AI. The survey also tied operational strain to fragmented tools and slow implementations.

Transaction monitoring was the most common AI use case. Globally, 30% of respondents said they use AI or machine learning for transaction monitoring, rising to 45% in APAC.

Integration gaps

The findings also point to a gap between partial and full integration of fraud and AML systems. While 95% reported "some integration," only 47% said they run fully integrated workflows. The rest rely on partial connections.

Respondents also highlighted the data impact of disconnected tools. In the survey, 80% said it is challenging to achieve a unified view of data. SEON argued this limits how much value organisations can extract from AI across fraud and compliance.

Implementation timelines were another friction point. Only 10% said they go live in under two weeks. Another 38% cited one to three months, while 24% said four months or longer. When roll-outs run long, 52% cited increased costs and 47% cited prolonged exposure to fraud.

Vendor strategy is also shifting: 85% plan to add a vendor and 49% plan to replace one. The survey did not specify which tool categories firms intend to add or retire, but the results suggest continued churn in the fraud and AML technology stack.

Headcount still rising

The report suggests the spread of AI agents has not reduced the perceived need for staff. In the survey, 85% said they view AI agents as support or augmentation rather than replacement. Only 12% expect eventual replacement.

Threat types remain varied across sectors. Account takeovers ranked as the top fraud threat at 26% globally and 24% in APAC. Promo and discount abuse came in at 18% globally and 26% in APAC, while return fraud was cited by 18% globally and 8% in APAC.

SEON linked the continued build-up of teams to the scale of activity organisations must manage, as well as structural issues such as siloed teams and disconnected datasets.

"Fraud and financial crime were supposed to become more manageable as AI matured," said Tamas Kadar, CEO and co-founder of SEON. "Instead, 2026 is the year leaders are confronting a more complicated reality. AI adoption is real, confidence is high, but the scale and pace of fraud - compounded by fragmented systems - continue to drive increased investment rather than reduced overhead. The bottleneck is no longer whether AI works. It's everything around it: disconnected data, siloed teams, slow implementations. The organisations that pull ahead will be the ones that unify fraud and AML intelligence, shorten the distance between threats and controls, and treat integration as strategy, not plumbing."

Governance focus

The report also points to a shift in priorities from adoption to governance. SEON said the conversation is moving from whether AI works to whether organisations can trust it, with greater attention on explainability and accountability.

In the survey, 78% of respondents said decentralised digital identity will become central to fraud and AML work. Data privacy regulation was cited by 33% as the biggest external force shaping AML, while 25% pointed to criminals' advancing use of AI and obfuscation techniques.

Across performance segments, SEON said organisations growing at 51% or more were nearly twice as likely as slower-growing peers to say achieving unified visibility is "not very challenging."