Case Study: BNPL provider achieves 41% reduction in false positives and 5x review efficiency with DataVisor

Raising the Bar for Customer Experience and Reducing Fraud with Machine Learning

BNPL, an innovative American fintech that pioneered point-of-sale lending, was facing synthetic identity fraud, account takeovers and promotion abuse that outgrew its insourced fraud model and forced a trade-off between customer experience and risk control. The company engaged DataVisor to deploy its fraud and risk platform—using the Feature Platform, Rules Engine, Knowledge Graph and Case Management tools backed by unsupervised and supervised machine learning—to detect coordinated attacks while preserving a smooth borrower experience.

DataVisor ingested BNPL’s data, automated feature engineering, deployed ML models and rule sets, and used link analysis and one‑click investigations to surface and act on complex fraud rings; the solution was integrated and delivering value within two weeks. The result for BNPL was a 41% reduction in hurt ratio (false positives), 320+ fraud rings detected (some tied to over $80K losses per attack) and an estimated 5x improvement in review efficiency, producing substantial cost savings and faster decisions.

BNPL