Insurers tackle fraud with technology

Virtually every enterprise is affected by losses due to fraud, but industry segments that process large numbers of transactions are particularly vulnerable. Every day, the number of electronic transactions continues to grow, increasing both the potential risk of fraud and the potential size of fraud losses. Analyzing the large volume of data from these transactions to detect fraud and abuse is simply beyond the ability of human beings, however skillful they may be. So to keep up with the times, insurers are increasingly turning to neural network technology. Neural network fraud-detection systems mark a significant jump ahead in the insurance industry's technological race against fraudsters, and may save insurers--and their clients--millions of dollars.

How Fraud Got Out of Control

Insurance fraud costs the average American household more than $5,000 a year in the form of higher premiums and prices for goods and services, according to Insurance Fraud: Renewing the Crusade, a recent study from Hartford, Connecticut-based Conning & Company. In 1999, consumers paid an estimated $96.2 billion in increased premiums and more than $530 million in the increased cost of goods and services--all thanks to fraud. The epidemic stems from various sources: