The existing claims fraud technology is reactive, and organized crime syndicates go undetected since each claim is viewed in isolation. Currently, most insurers have only implemented business rules and red-flagging technology based on past claims experience. Often, when fraudulent activities are detected, insurers find that pursuing the perpetrators is not cost effective, and they simply accept it as another cost of doing business.
There are a number of different technologies that insurers are beginning to use to tackle this growing problem, including predictive analytics, network link analysis and voice stress recognition software. Sophisticated analytics help insurers to analyze structured and unstructured claims data to discover previously undetected claims fraud. Network link analysis has proven effective in identifying organized fraud activities by modeling relationships between entities in claims. Although somewhat controversial, stress-recognition software is being used by insurers to assess the stress levels of a claimant during interviews and customer service calls for the likelihood that the individual is lying.
Insurance companies are using analytics to build programs that produce fraud propensity scores. Once data is entered, the claim is automatically scored for the likelihood to be fraudulent and, if necessary, is made available for review by the SIU team. Insurance companies need to implement an online, real-time claims strategy based on a combination of fraud-detection technologies to prevent fraud before it happens. Also, insurers should implement fraud propensity scoring in the future within their new-business process to deny coverage on policies that have a high tendency for fraudulent activities.