Case Study: Prominent North American Bank achieves more predictive, explainable credit risk decisions with FICO Analytics Workbench

A top 5 US bank leverages machine learning to responsibly improve credit risk decisions

A top‑five U.S. bank with some 70 million customers sought to bring advanced machine learning into its consumer credit risk program while meeting strict regulatory requirements for transparency, fairness, and governance. The challenge was to move beyond traditional scorecards across multiple lending lines (cards, auto, mortgages, personal and student loans, small business) and prove that ML could deliver better, explainable decisions without sacrificing auditability or legal defensibility.

The bank adopted FICO Analytics Workbench with FICO’s Explainable AI to combine scorecard and ML development in a single, cloud‑deployed environment. A side‑by‑side validation showed ML models delivered a measurable performance edge (notably for delinquent subpopulations), used richer data dimensions, and remained transparent and controllable for legal and regulatory review. The deployment delivered increased predictive power, faster model deployment, improved data‑science productivity and collaboration, and the ability to scale ML safely into production.