BUSINESS SITUATION
- Our client had a legacy fraud management system with 100s of rules stacked over a period plus traditional models leading to high false positive rates and customer dissatisfaction
- Using our AI/ML-driven models and strategy rules, we were able to help our client prevent massive fraud losses and reduce false positive rates
SGA APPROACH
- We created intelligent features using the historical transaction data and accumulated it
- We created futuristic-kind data-driven segments and built models for each segment
- A closed loop fraud model strategy was implemented to detect outliers and identify emerging anomalous patterns in the transactions made by the customers
- We implemented advanced AI and self-learning models enabling us to reduce false positives
ENGAGEMENT
Robust fraud models to reduce false positive & increase fraud detection rate for a leading US credit card issuer
BENEFITS AND OUTCOMES
- Reduced decline rates by 100-150 bps resulting in £16M-20M incremental turnover
- Over 40% reduction in false positive rates
KEY TAKEAWAYS
- Significant improvement in both high-level transaction fraud detections and false positive reductions
- Improved payment acceptance rates leveraging card acquisition data
- 40% increase in fraud capture
- 70% decrease in hold rates
- 50% reduction in sales declines
