Case Study: Leading U.S. Credit Card Issuer captures 25% more application fraud and saves $15M+ with DataVisor’s machine learning solution

Leading U.S. Credit Card Issuer Uses DataVisor’s Machine Learning Solution to Reduce Application Fraud Losses

Leading U.S. Credit Card Issuer, which processes over 20 million applications a year, was facing rising waves of coordinated third‑party and synthetic application fraud that existing rule‑based and supervised models could not catch. High false positives and large alert volumes were delaying legitimate applicants and driving up manual review costs, so the issuer turned to DataVisor and its unsupervised machine learning solution to shore up real‑time detection and reduce operational strain.

DataVisor integrated its unsupervised ML engine (leveraging its Global Intelligence Network) within weeks without requiring labeled data, surfacing linked fraud rings, enabling bulk decisions, and accelerating model iteration. The deployment delivered measurable impact: 25% additional fraud captured, 94% detection accuracy, a 0.17% good‑user false positive rate, and more than $15M in combined fraud loss and operational savings, while catching attacks days earlier.