Elizabeth McCaul, former member of the European Central Bank’s Supervisory Board, recently addressed the European Anti-Financial Crime Summit 2026, focusing on how digital twin technology and artificial intelligence are reshaping compliance in the financial sector. Her speech emphasized the shift from traditional rule-based anti-money laundering (AML) systems to performance-driven models that use advanced data analytics and simulation.
McCaul reflected on her experience during the aftermath of the September 11 attacks when she was New York’s Superintendent of Banks. At that time, financial crime compliance systems were basic and heavily reliant on broad, rule-based filters that generated many false positives. These early systems were foundational but limited in their ability to detect complex criminal behaviors.
Over the past 25 years, AML systems have evolved significantly. Initially, transaction monitoring was isolated and static, relying on fixed thresholds and manual reviews. By the 2010s, institutions began incorporating analytics like network analysis and machine learning to enhance detection. However, a gap remains between how institutions describe their controls and how effectively those controls work in practice.
The new EU Anti-Money Laundering Authority (AMLA) regulations highlight this challenge by demanding not just evidence of controls but proof of their effectiveness in real time. AMLA encourages financial institutions to move beyond periodic reviews toward continuous performance measurement and improvement. This approach requires detailed visibility into what risks are detected, missed, or evolving.
McCaul introduced digital twin environments as a promising solution to this problem. A digital twin replicates an institution’s existing AML systems in a controlled simulation space where AI can optimize detection rules without affecting live operations. This method allows banks to test and improve their systems continuously by replaying historical transactions, identifying gaps, and simulating new risk scenarios.
Agentic AI takes this concept further by autonomously generating and testing new detection hypotheses, enabling institutions to adapt quickly to emerging threats without waiting for manual updates. This technology layer can be applied over legacy systems, preserving past investments while significantly enhancing performance.
Drawing parallels with mobile banking innovations in Africa, McCaul argued that financial institutions do not need to replace costly legacy systems entirely but can leapfrog limitations by adding intelligent overlays. This strategy supports a shift from compliance based on frameworks toward measurable outcomes backed by data and simulations.
Ultimately, McCaul sees these technological advances as vital for meeting AMLA’s requirements and improving financial crime prevention globally. Institutions adopting digital twins and AI-driven optimization will be better equipped to demonstrate their effectiveness to supervisors and law enforcement, turning compliance into a strategic advantage rather than just a regulatory burden.
