Despite the fact that insurers have started to put more resources into fraud-mitigation technology, Light asserts, the traditional "red flag" method -- in which adjusters use predetermined lists of about 30 circumstances to gauge if a claim may be fraudulent -- is still over-used. "Red flags are one of the more traditional or legacy methods of identifying fraud," he says. "The problem is they rely totally on experience and distilled best practices. They're limited in terms of being able to scan the totality of the claims and discover complete sets of fraud methods."
Light says that a small minority of adjusters currently reports a majority of fraudulent claims to SIUs (special investigations units), a sure sign of inconsistency. "You'd expect adjusters who are working on the same kinds of claims to have roughly similar incidents of fraud, and therefore [they] should make roughly similar sets of referrals" he explains.
Meanwhile, several newer techniques -- including predictive modeling, neural networks, profiling and identity matching -- have gone underutilized, Light notes. "These are all established technologies that have been around for years," he explains. "Some have been used in other parts of the insurance enterprise but only fairly recently have been brought to bear on fraud." Light contends that insurance lags behind the credit card and banking industries regarding fraud technology adoption.
Know When to Investigate
One of the most difficult aspects of insurance fraud mitigation is knowing when to refer a claim to the SIU and when not to. Too few referrals and fraudulent claims go unchecked; too many and lawsuits and poor public relations problems can persist.
In part by employing business rules and workflow design, the newer technologies can create more-accurate fraud scores that go beyond the 20 or 30 red flags with which a typical claims adjuster would be equipped, according to Light. He explains that more-granular fraud scores at certain levels would trigger deeper investigation or SIU referral.
"There are a lot of technologies out there that can make the identification of potential fraud much more sophisticated and fine-grained, and can make the subsequent investigation of those fraudulent claims much more effective through the application of more-powerful tools," Light says.