Fraud Detection Platform

Leading Bank in Southeast Europe

Business Challenge​

Banking information systems are increasingly exposed to fraudulent financial transactions. The key in protection from fraud is to reduce the time between occurrence and detection of a fraudulent transaction.


Our Approach​

Cloudera big data platform and Kafka message broker were implemented to continuously monitor transactions and ensure the detection of the fraudulent activity in real time. A set of rules based on Drools business rule management system was defined to recognize and flag potential fraud. Marked transactions are reported to the control system in the shortest time possible to stop execution and analyse causes.



System that applies set of rules for fraud detection to all transactions as soon as possible and mark these transactions as a potential fraud. The appearance of such marked transaction are then reported to the control system in shortest time possible in order to stop further transaction execution and initiate action to detect the cause.