Insurers depend on prebuilt business rules, workflow, and information collected from customers to detect fraud in claims. Claims handlers must refer to long lists of fraud red flags that need to be checked against the specifics of the cases under management, which slows down the claims handling process overall, with high percentages of false positives and unknown numbers of false negatives.
Goals and Objectives
Use multiple internal and external data sources as well as image analysis tools to verify the reliability of the data provided and to predictive models aimed at flagging the 2% to 3% of cases with high fraud-risk and allow for straight-through processing of the remaining 97% to 98% non-suspect cases.
IoT, Advanced analytics, Cognitive technologies, Cloud, Fraud visualization, fraud ring identification, Document forgery detection