Teradata Korea announced that Danske Bank has launched an AI-based fraud detection platform in collaboration with Teradata’s Think Big Analytics.

The platform's engine analyzes tens of thousands of latent features through machine learning and scores millions of online banking transactions in real time to provide actionable insights into all fraudulent activities. Danske Bank has significantly reduced the cost of investigating false positives, which are the misidentification of normal transactions as 'fraud'.

Danske Bank's existing fraud detection system had to apply manual rules more extensively over time. This increased the cost and time required to investigate a massive volume of false positives, amounting to 99.5% of all transactions. Meanwhile, the bank's large fraud detection team faced problems of increased workload and decreased utilization.

From a modeling perspective, fraud cases are still extremely rare, occurring at a rate of 1 in 100,000. The team reduced false positives in this model by 50% and increased the actual detection rate to approximately 60%. Danske Bank's fraud prevention program is cited as the first instance of applying machine learning techniques to the field of operations and developing a deep learning model to verify these techniques.

Mads Ingwar, Director of Customer Service at Think Big Analytics, stated, "Every bank needs a digitalization roadmap and strategy to introduce data science into the organization, including scalable, advanced analytics platforms. For online transactions, credit cards, and mobile payments, banks require real-time solutions. The cutting-edge AI-based fraud detection platform, jointly developed with Danske Bank, scores transactions in less than 300 milliseconds. This means that when a customer purchases groceries at a supermarket, the system scores the transaction in real time and provides immediate actionable insights. This type of solution is a first for the financial services industry."