A 2023 SAS consumer fraud study found that roughly 70% of people had been victims of fraud at least once, and nearly 40% had experienced multiple incidents. These figures underscore the urgency for financial institutions to upgrade their fraud-prevention strategies.

“Fraud has become increasingly sophisticated, posing substantial risks to financial institutions,” Maric explains, noting that social engineering tactics leave little time for intervention. She observes that fraudsters have readily adopted AI tools, using generative AI to create deepfake videos, synthetic identities and highly personalised phishing scams. Analysts project that generative-AI-driven fraud losses could rise from $12.3 billion in 2023 to $40 billion by 2027. A SAS-ACFE report found that 83% of fraud professionals plan to deploy generative AI in their defences within two years.

Traditional fraud management approaches are now under strain. Maric highlights several challenges: data often remains siloed across departments and channels, legacy rules cannot adapt quickly to new schemes, and the rise of instant payments increases exposure to loss.

A SAS roundtable also identified ‘managing siloed data’ and ‘implementing real-time fraud detection’ as top challenges. Without fast adaptation, firms endure high false-positive rates and frustrated customers.

Maric stresses that fraud prevention frameworks must be flexible and continuously updated to keep pace with threats. She and her team advocate a holistic, AI-driven strategy. A hybrid AI approach that combines machine learning (supervised and unsupervised), advanced rules and network analytics is ideal to spot fraud at multiple levels.

“Profiling devices, applications and customers for transactions can improve fraud detection and reduce false positives,” Maric says. “Integrating siloed data, such as device IDs and IP addresses, into core banking systems provides a unified view of risk. Leveraging such integrated analytics allows fraud teams to spot subtle anomalies that legacy systems would miss.”

Maric also highlights the use of synthetic data in model development. Testing models on generated fraud data helps banks prepare for novel attacks.

“Developing and testing models with synthetic data can improve resilience to unknown threats,” Maric notes.

This forward-looking practice builds adaptability into fraud prevention frameworks, which is at the heart of the Fraud Prevention 2.0 approach.

Signs of change are already evident. A global FT/Longitude survey (conducted with SAS) found that 75% of banks plan to increase investment in risk-technology infrastructure. This significant rise (from 51% in 2021) underscores how financial leaders are prioritising analytics and AI to build resilience. SAS and its partners report that institutions are shifting from reactive fraud detection toward proactive, predictive strategies that combine AI, automation and continuous monitoring to catch fraud before it occurs.

SAS continues to drive thought leadership in financial crime risk management. The company emphasises not only technology but also awareness and collaboration. For example, one key takeaway from recent SAS forums is the importance of customer education: financial firms should use awareness campaigns, training videos and gamified learning to help users spot scams.

Internally, Maric advises that fraud, compliance and cybersecurity teams share threat intelligence closely.

“Collaboration amplifies the value of individual efforts and ensures a cohesive defence,” Maric says.