Case Study: Danske Bank achieves 60% reduction in false positives and 50% increase in fraud detection with Teradata

Danske Bank Fights Fraud with Deep Learning and AI

Danske Bank, a Nordic universal bank serving Denmark, Finland, Norway and Sweden, faced growing fraud risk as customer interactions moved online and fraud schemes became more sophisticated. Its legacy rules-based system detected only about 40% of fraud, generated up to 1,200 false positives per day, and required investigators to chase many non-fraud cases, so the bank needed a modern analytics approach.

Working with Teradata, Danske Bank deployed an enterprise solution using deep learning, GPU-optimized infrastructure, integrated data pipelines and a champion/challenger model framework to make real-time, AI-driven decisions with human oversight. The rollout (live in five months) cut false positives by 60% (with expectations up to 80%), increased true positives by 50%, freed investigative resources for real fraud, and the platform is now being applied across other parts of the bank.

Danske Bank