Fuzzy Payment Screening Implementation at VP Bank Group

This case study details the implementation of fuzzy payment screening using machine learning at VP Bank Group. The primary objective was to update sanctions list screening in payment transactions by replacing exact searches with fuzzy matching. The project involved calibrating the system through effectiveness tests conducted by external consultants and efficiency testing based on historical transaction data, spanning a duration of seven months. The results indicated a clear improvement in the quality of hits and a significant reduction in the workload for the payments staff. The new system, developed in collaboration with ACTICO, aimed to enhance the identification of risky transactions while minimizing false hits. The study outlines the challenges faced, including the trade-off between effectiveness and efficiency in payment screening. The findings suggest that the implementation of machine learning techniques has led to a steady decline in the volume of hits while improving their quality, thereby optimizing the payment screening process.