SEON Report: AI Coexists With Growing Fraud Teams
Despite near-universal AI adoption, corporate fraud and compliance teams are still growing in size and budget, a new SEON report finds. The study, which surveyed over 1,000 global leaders, suggests that AI is augmenting rather than replacing human roles in these departments. The report also highlights challenges with fragmented systems.
- According to the SEON report, 83% of fraud and compliance leaders expect their budgets to increase in 2026, and 94% plan to increase headcount, up from 88% the previous year. The top use case for AI is currently transaction monitoring, cited by 30% of leaders. - The growth in fraud teams is partly a response to the increasing sophistication of criminals who are using generative AI to create realistic deepfakes, synthetic identities, and hyper-realistic phishing scams. Global deepfake fraud reportedly rose 700% in the first quarter of 2025. - A primary challenge limiting AI's effectiveness is data fragmentation; 80% of leaders say getting a unified view of data is a challenge. Siloed data across legacy systems prevents AI models from detecting suspicious patterns effectively and slows down investigations. - Only 47% of companies run fully integrated workflows between their fraud and Anti-Money Laundering (AML) systems, with the rest relying on partial or disconnected connections. This fragmentation leads to operational inefficiencies and an incomplete view of customer risk. - The prevailing view is that AI acts as a "digital coworker," augmenting human teams rather than replacing them. AI automates high-volume, repetitive tasks, which allows human professionals to focus on more complex investigations and nuanced cases that require empathy and critical judgment. - The conversation around AI in fraud prevention is shifting from whether it works to questions of trust, governance, and accountability. Key future concerns include the impact of data privacy regulations like GDPR (cited by 33% of leaders) and criminals' advancing use of AI (cited by 25%).
