@First Tool Helps MetLife Auto & Home Fight Fraud

A recent Accenture study reports that the majority of Americans believe that insurance fraud occurs because people think they can get away with it. As part of an effort to combat this perception while detecting the practice, carriers like MetLife Auto & Home, a subsidiary of MetLife (New York, $350.2 billion in assets under management), have taken "an aggressive stance on fraud supported by a field-based approach with investigators stationed throughout the U.S.," explained John Sargent, special investigations unit, manager, MetLife Auto & Home, at the IASA session titled "Waging the War on Fraud." The conference took place last week in Las Vegas.

Although investigators are a crucial line of defense in the war on fraud, cases first must be referred to them, and the sooner initial indicators of suspicious activities are detected, the better. In an effort to protect its organization further from vulnerabilities associated with fraud exposure at the first notice of loss, Sargent explained that MetLife Auto & Home teamed up with El Segundo, Calif.-based CSC in 2002 to develop @First, an automated first notice of loss fraud detection tool that uses a combination of predictive data modeling technology, company claims data and industry data sources to help the underwriter flag suspicious claims.

"We've taken five years worth of claims information and applied link analysis techniques to it," Sargent explained. "The method gives us a model of what to look for so that we can detect whether a policyholder's claim [is suspicious] based on prior claims history." With the help of CSC, MetLife Auto & Home has also developed three rules engines that work with @First. Sargent related that the carrier's business rules are applied to industry data sources. For instance, if no police report can be located for a corresponding accident that a claim has been filed for, that claim is flagged as suspicious.

So far, the carrier has increased its referral rates to investigators by 22 percent. Still, Sargent conceded that @First cannot filter all cases of fraud. "Although the system helps us to detect fraud, it's not good at helping to detect first time offenders," Sargent admitted. Still, as the carrier improves its business rules, Sargent expects that it will be able to serve policyholders filing legitimate claims better. "Although 10 to 20 percent of claims may be fraudulent in some way, there is a chance that 90 percent of claims are legitimate," Sargent related. "We owe it to those policyholders to process their claims fairly and expeditiously."