This report is based on insights from a survey commissioned by SEON, gathering perspectives from fraud prevention and compliance professionals across industries. Research and analysis were conducted by Jane Singh & Katy Chrisler. The findings are based on a quantitative third-party survey of 1,010 fraud, risk and compliance leaders working in digital-first organizations, all holding director-level or more senior roles.

AI Reality Check: 2026 Fraud & AML Leaders Report reveals an industry that has fully embraced AI and is discovering that the hard work is only just beginning. AI is real, embedded and widely trusted, but it has not materially reduced the scope of fraud and AML operations. In many cases, criminals’ use of AI has stretched that scope faster than efficiency gains can keep up.

One signal stands out in the year-on-year data. Last year, 56% of leaders disagreed or strongly disagreed with the statement, “fraud losses are growing faster than our revenue;” this year, that figure has dropped to 35%, a shift that suggests many organizations now feel losses pressing closer to, or even outpacing, growth. As AI becomes a tool for both defenders and attackers, simply keeping up has become a larger, not smaller, task.

Key Takeaways

1. AI is everywhere, but simplification isn’t

• AI is now baseline infrastructure, not a pilot. Ninety-eight percent of leaders report that their teams are already integrating AI into day-to-day workflows, with AI/ML for transaction monitoring being the most common and mature use case (30%).
• Confidence is high, but pressure is higher. Ninety-five percent are confident that AI can reliably detect and prevent fraud, and 52% are very confident — yet most report an increase in volume, structural complexity and regulatory and operational scrutiny, rather than a decrease.

2. Investment keeps rising across budgets, headcount and tools

• Budgets continue to grow. Eighty-three percent of organizations expect their fraud and AML budgets to increase in 2026, only slightly below last year’s 86%, signaling that modernization is a strategic priority.
• Hiring is accelerating, not shrinking. Ninety-four percent of leaders plan to add at least one full-time fraud/AML hire in 2026, up from 88% the prior year, with one-third planning 3–5 hires, another third planning 6–10 and 17% planning to add more than 10 roles.

3. Automation is redefining teams, not replacing them

• AI agents are seen as copilots, not substitutes. Over 85% of leaders believe AI agents should support or augment analysts, while only 12% think they should eventually replace them.
• Human roles are shifting up the value chain. As automation absorbs volume and routine triage, teams are reallocating time toward complex investigations, model oversight, regulatory reporting and cross-functional strategy.

4. Fragmentation is the new bottleneck

• Integration is widespread on paper, brittle in practice. While 95% of organizations report at least some integration between fraud and AML workflows, only 47% run fully integrated platform workflows; the rest rely on partially connected systems that create friction and blind spots.
• Unified visibility remains elusive. Eighty percent say it is at least somewhat challenging to obtain a unified view of data and insights across fraud and AML systems, and more than 40% describe it as extremely or very challenging, undermining both AI performance and operational efficiency.

5. Time-to-value compounds every other risk

• Go-lives still take months. Most organizations report that implementing a new fraud or AML solution takes multiple months after vendor selection. Specifically, 38% go live within one to three months, while another 24% require four months or more.
• Delays have real costs. When implementations overrun, leaders most often cite increased operational costs (521 respondents), prolonged exposure to fraud risks (475) and sustained manual workarounds reflected in higher operational load and technical effort as the top consequences, not simply IT inconvenience.

6. High-growth leaders look different

• Growth and integration move together. High-growth organizations (26–50% and 51%+ revenue) are more likely to be very confident in AI and almost twice as likely as slower-growing peers to report lower levels of difficulty achieving unified visibility, likely reflecting earlier and more sustained investment in integration.
• They treat integration as a strategy, not plumbing. Compared to slower growing peers, high-growth organizations are more likely to unify fraud and AML intelligence, and the data suggests they are also more inclined to reduce reliance on fragmented point solutions and prioritize shared data foundations.

7. The next frontier: governance, identity and skills

• Leaders are increasingly focusing on explainability, auditability and human accountability as AI becomes embedded in high-stakes risk decisions.
• Decentralized identity is on the radar. Eighty-five percent of respondents believe decentralized digital identity will become part of the future fraud and AML stack, even as timelines remain uncertain and regulatory clarity continues to evolve.
• Skills are evolving toward AI, data and collaboration. Fraud teams are prioritizing AI, machine learning and advanced data analytics, alongside cross-departmental collaboration and customer experience optimization, as the most critical future capabilities.

Methodology & Respondent Profile

This report presents a global, leadership-level perspective on the evolution of fraud and AML in the era of AI. It focuses on what leaders are actually doing and measuring today so that investment decisions can rest on evidence rather than intuition.

The findings are based on a quantitative third-party survey of 1,010 fraud, risk and compliance leaders working in digital-first organizations. All respondents hold director level or more senior roles, with direct responsibility for fraud prevention, AML compliance, risk management or adjacent decision-making, ensuring the insights reflect how strategy and operations actually meet in practice, not just theoretical opinions.

The sample is intentionally global, with substantial representation from North America (NOAM); Europe, the Middle East and Africa (EMEA); Latin America (LATAM) and Asia Pacific (APAC), capturing a mix of both mature and fast-growing fraud environments and highlighting how regional nuances shape threats, regulation and technology choices. The respondents are distributed across North America (263), EMEA (252), LATAM (243) and APAC (252), as well as across payments/fintech/financial services (347), retail/eCommerce (331) and betting and gaming (332).

2026 in Context

Nearly 80% report revenue growth above 10% in the past 12 months, with 48% in the 11–25% range, 36% in the 26–50% range and 4% growing at 51%+. In that environment, fraud and AML teams sit directly between growth ambitions and risk exposure.

This combination of product and geographic expansion fundamentally reshapes the fraud landscape. It drives higher transaction volumes, introduces new payment methods and behaviors and increases regulatory exposure across multiple jurisdictions, all of which expand the surface area fraud and AML teams must monitor and control.

Ninety-seven percent of companies plan a major growth move in 2026.

• Account takeovers: 26% — the most commonly cited threat, reflecting rising credential-based attacks
• Promotion or discount abuse: 18% — frequent exploitation of marketing offers and incentives
• Return fraud: 18% — persistent challenge linked to lenient refund policies
• Chargebacks: 16% — often tied to disputed transactions or false claims
• Loyalty or rewards program abuse: 13% — targeting stored value and points-based systems
• Alternative payments abuse (e.g., crypto): 6% — reflecting the growing risk in emerging payment channels
• First-party or “friendly” fraud: 4% — lower in volume but difficult to detect and dispute

AI Has Moved From Experimentation to Expectation

With 98% of leaders already integrating AI into their workflows, just 2% are still in the planning phase. This adoption spans regions and verticals, embedding AI into how alerts are triaged, cases are investigated and risk decisions are made.

The deepest penetration is in core monitoring and scoring. AI & ML for transaction monitoring tops the list at 30%, making it the most common and mature application. AI risk scores with explainable outcomes are used by 14%, AI-generated summaries for alerts and cases are live at 12%, and AI & ML for anomaly detection are in place at 11%. Almost no one is sitting out entirely: 2% are not currently using AI but plan to, and 0% have neither used nor intend to use AI for fraud and AML purposes.

Specialized capabilities are emerging at more minor, but meaningful, scales: AI regulatory reporting/automatic form fills are used by 13% of organizations, AI SAR narrative generation and AI agents for sanctions screening each achieve a 5% success rate, and AI agents for transaction monitoring are used by 8%, typically as copilots that recommend actions rather than fully automate them.

Confidence in AI is high. Ninety-five percent of leaders are somewhat confident that AI solutions can reliably detect and prevent fraud, and more than half (52%) describe themselves as very confident. Just 4% are “not very confident” and 1% “not confident at all.”

This confidence rises with growth. Among organizations projecting 0–10% revenue growth, 40% are very confident in AI; this increases to 48% for those expecting 11–25% growth, 59% for the 26–50% group and 68% among those forecasting 51% or more increase. Faster-growing companies are more likely to see AI as a dependable part of their risk stack rather than a test bed.

Investment Keeps Climbing

Looking ahead to 2026, 83% of organizations expect their fraud and AML compliance budget to increase, while only 17% expect it to remain flat or decrease. This follows a period in which 86% said their fraud budget would increase, indicating that elevated spend is becoming the norm rather than a temporary spike.

On the people side, the share of organizations planning to add at least one full-time fraud or AML hire has risen from 88% in 2025 to 94% in 2026. Within that 94%: 11% plan to add 1–2 full-time hires, 33% plan 3–5 hires, another 33% plan 6–10 hires, 17% plan to add more than 10 full-time roles, 3% plan to add contractors and only 3% report no plans to increase headcount.

On the technology side, almost everyone expects to touch their stack: 85% plan to add a vendor, 49% plan to replace a vendor, 20% plan to remove a vendor and just 3% expect no significant updates.

Automation Is Not Reducing Headcount: It’s Redefining It

Automation has not led to leaner fraud and AML teams; it has coincided with more hiring. For the 2025 planning cycle, 88% of leaders expected to increase headcount. For 2026, that share has risen to 94%, confirming that staffing continues to increase, not decrease.

What is changing is the nature of the work. Rather than simply adding more analysts to do the same tasks, leaders are orienting roles toward higher-judgment activities that cannot be fully automated — creating demand for people who can bridge systems, reconcile signals and align teams across fraud, AML, product, engineering and compliance.

Views on AI agents crystallize the new division of labor. Only 12% say agents should eventually replace analyst tasks; 40% believe agents should support analysts with recommendations and summarization based on standard operating procedures; 38% say agents should augment investigators, providing a starting point for deeper investigations, but not replacing them; 7% see agents as promising but not yet mature; 3% prefer human-driven decisions with minimal AI involvement; and 1% say it is too early to tell. In total, over 80% of leaders place AI agents in a supporting or augmenting role, rather than a replacement role.

What the Data Is Really Telling Us

At first glance, the numbers appear contradictory: AI usage is almost universal and confidence is high, yet budgets and headcount continue to rise. AI has improved detection at the workflow level, but scale, fragmentation and oversight requirements have expanded the overall workload faster than efficiency gains can offset it.

Three structural forces emerge from the data as the main drivers of simultaneous automation and hiring:

• Volume is scaling faster than efficiency gains. Around 80% of companies have increased revenues by more than 10% in the past year, translating into more users, sessions and transactions across more products and geographies.
• Oversight and governance requirements are rising. Regulators emphasize explainability, auditability and human accountability, raising demand for skilled humans in model monitoring, validation, exception handling and regulatory reporting.
• Fraud threats are adapting alongside defenses. New fraud patterns emerge more quickly and static controls become increasingly ineffective, keeping human judgment essential even as automation accelerates detection and response.

Fragmentation Becomes the Bottleneck at Scale

Ninety-five percent report at least some integration between fraud prevention and AML workflows, yet only 47% operate a fully integrated platform or workflow. Another 47% rely on partial integration — some shared data, but separate systems — and 6% run separate or non-existent integrations (5% with little coordination, 1% with none). This “almost integrated” state becomes fragile at scale: as transaction volumes, products and markets grow, seams between systems become more visible and more costly.

In total, 80% of leaders experience at least some difficulty achieving a single view, and more than 40% are at the “extremely” or “very” challenging end of the spectrum. These data silos create operational tax: slower investigations, more false positives, duplicated effort and greater reliance on manual reconciliation between fraud and AML teams.

Growth brings sharper consequences for fragmentation. Only 19% of all respondents say unified visibility is not very challenging or not challenging at all — but that share rises with revenue growth: 23% among companies with 0–10% revenue growth, 16% for 11–25% growth, 20% for 26–50% growth and 39% for the 51%+ growth group. Mid-growth companies often feel the sharpest pain: they have outgrown ad hoc integrations but have not yet fully invested in unified architectures.

Time-to-Value Remains Sluggish

Even as cloud and AI tools mature, implementation speed has not kept pace with business urgency. After selecting their most recent fraud or AML vendor, 38% of organizations go live within one to three months, while 18% take four to six months, 5% take seven to twelve months and 1% take more than a year. Only 10% go live within two weeks.

When timelines slip, the impact is tangible. Among respondents whose go-live took longer than planned, the most common adverse effects were increased operational costs (52%), prolonged exposure to fraud risks (47%), required excessive technical resources (36%), negative impact on team productivity (32%), frustration among stakeholders (27%), delayed realization of ROI (26%) and diverted focus from other business opportunities (25%). Only 14% said the impact was “minimal,” and 7% indicated their go-live did not take longer than planned.

What Leaders Are Watching Next

With 98% of organizations already using AI in fraud and AML, and 95% expressing at least some confidence in its reliability, the conversation is shifting from basic effectiveness to governance, explainability and control. The integration and visibility challenges — where 80% find unified views at least somewhat difficult — reinforce why this governance lens matters: trust in AI must extend beyond individual scores to the entire decision pipeline.

Beyond AI, leaders are watching shifts in identity and regulation. Asked whether decentralized digital identity (e.g., Web3 verifiable credentials) will become a significant part of fraud and AML prevention: 78% say yes — it will be central to the future; 6% say no; 15% say it is too soon to tell; and fewer than 1% say they don’t know enough to answer.

On the broader regulatory and threat landscape, one in three leaders (33%) points to data privacy regulations (e.g., GDPR, CCPA) as the single external factor expected to have the biggest impact on AML compliance over the next two years, followed by criminals’ advancing use of AI and obfuscation techniques (25%), growth of crypto/decentralized finance adoption (11%), expanding sanctions regimes (10%), political/regulatory uncertainty (9%), increasing cross-border transactions (9%) and shortages in skilled AML compliance talent (5%).

As these trends converge, leaders are increasingly prioritizing skills in AI, machine learning and advanced data analytics; system architecture and technology integration; regulatory and compliance knowledge; and cross-departmental collaboration and communication. Together with the strong preference for AI agents as support rather than replacement, these priorities point toward teams that act as designers and governors of intelligence, rather than simply operators of tools.

Final Reflections: Where Leaders Go From Here

The numbers here mark a clear inflection point. AI now runs through almost every fraud and AML program, yet leaders still add personnel, increase budgets and struggle with fragmented systems. The old assumption — that automation would simplify operations — no longer holds.

1. From tools to architecture. Most organizations already own enough point solutions. The real advantage lies in connectivity — whether fraud and AML share a data backbone, whether signals travel across products and regions and whether decision logic stays explainable end-to-end. The gap between the 47% with integrated workflows and the 80% who can’t view everything in one place speaks volumes.
2. From projects to muscle memory. Time-to-value is now a structural property of the organization, not a project milestone. Leaders who deploy new vendors and controls in weeks rather than quarters narrow their exposure window every time they move.
3. From operators to system designers. As AI takes on pattern-matching and triage, fraud and AML roles shift toward pattern-setting — treating rules, models, integrations and KPIs as design choices to iterate on, not fixed constraints.

Related resources

• The Total Opportunity Cost of Fraud Is Bigger Than You Think
• Leveraging AI and Innovative Tech to Combat Emerging Complex Fraud Schemes
• Traditional Fraud Prevention Measures Aren’t Enough