Fighting fraud with technology: carriers need to equip themselves with integrated tools and technologies to ward off fraud.

Many carriers are struggling to combat the growing challenge of fraud. Aging technology and inefficient processes hinder the detection and prevention of fraudulent claims, resulting in higher costs for insurers, businesses and consumers.

Property/casualty insurance fraud cost insurers about $ 31 billion in 2002, according to the Insurance Information Institute. A recent Accenture survey found that more than one in five U.S. adults believes it's acceptable to overstate the value of their claims. A majority believe that insurance fraud occurs because people believe they won't get caught.

Another particularly distressing survey finding is that nearly one-half of the respondents who knew someone submitting a claim for an amount higher than the actual loss said that a doctor, auto body shop, insurance appraiser or other third party was involved in the fraud.

Insurers appear to have few allies in their battle against fraud. Indeed, one-third of survey respondents told Accenture that they were unlikely to report someone who had committed insurance fraud.

The key to fighting fraud is the rapid evaluation of large amounts of dynamic information. Many carriers, unfortunately, typically check demographic data only one time--when the claim is first filed. This practice opens the door for potential fraud.

A better approach is to implement Web-based technology such as "electronic agents" that continuously re-evaluate claims for fraud detection throughout the claim and alert the appropriate people when thresholds are exceeded or certain parties are involved.

For example, each time a new individual is introduced to the claim or changes in claim or demographic data occur (such as name, address, phone number, attorney, witnesses, physicians, medical condition), electronic agent technology rims the data against the carrier's predetermined fraud pattern and, when appropriate, sends potential fraud red flags to the claims adjuster's desktop for further investigation.

Consider this common fraud scenario: A claimant submits a seemingly authentic claim for an automobile accident injury but delays providing the claims handler with his attorney's name. Knowing that the carrier stops automatically checking for fraud 60 days after claim submission (despite the fact that Central Index Bureau members are required to reindex injury claimants every six months), the perpetrator submits the name of his attorney--who is engaged in the fraud ring--after 60 days. After that time, it's difficult for the carrier to manually piece together the evidence linking the claimant and the attorney. The electronic agent would be programmed to check the introduction of this new person to the file--no matter when it occurs--and to notify the handler any time the attorney's name appears in the file.

To fully leverage the electronic agent, carriers must build and maintain a sophisticated rules library. As the source of the fraud patterns against which the agent continually searches, the library should not only leverage databases from industry groups but also institutionalize a robust set of red flags based upon the carrier's claim experience. Examples of such red flags might include:

* nonwitnessed, single car accidents;

* history of multiple claims returned on the Central Index Bureau;

* claimant is difficult to contact or cancels appointments;

* claim amounts that are disproportionate to the injury or damage sustained; and

* seasonal workers claiming on-the-job injury as the season is about to end.

Unfortunately, many carriers still rely entirely on adjusters to manually indicate patterns that will trigger a referral to the special investigations unit. Ill-trained adjusters may fail to recognize the pattern or forget to check the appropriate box on the form and not make the referral. If adjusters make the referral, they're typically removed from the fraud investigation process, which may lead to duplication of efforts by claims and the SIU or, at the very least, inhibit the two units from working seamlessly together.

Automatic referral systems foster a collaborative relationship between claims and the fraud unit. Automatic referrals to the SIU based upon claim patterns are essential to early fraud detection. Claim pattern analysis scrutinizes incoming and current claims according to thousands of claim attributes. It then segments claims according to complexity, risk and other profiles that the carrier defines. The constant monitoring of claims for established patterns quickly identifies those that need special handling or that have a sudden change that calls for more expert attention. By automating critical aspects of the claims management system, technology can help insurers better address fraud.

Michael A. Lucarini is a partner with Accenture Insurance Practice. He can be reached at [e-mail].