BoE and London BIS’s Project Hertha tested AI to detect FC in real-time retail payments, boosting illicit pattern detection by 26% and account identification by 12%, while revealing key limitations

The project focused on identifying illicit finance patterns in real-time retail payment systems through advanced transaction analytics. Notably, it achieved a 26% improvement in detecting previously unseen criminal behaviours and a 12% increase in the identification of illicit accounts compared to conventional methods. These findings suggest that AI-driven analytics could serve as a powerful supplementary tool for banks and payment…