Insurers are more frequently using technology-based tools to detect insurance fraud, but only a minority are using technology to help identify non-claims functions such as underwriting fraud, a study commissioned by SAS and the U.S.-based Coalition Against Insurance Fraud has found.
The online survey of 74 insurers showed that 88% of respondents said they are currently employing anti-fraud technology, and nearly all said they used it for claims fraud detection and investigation.
The top three anti-fraud technologies used by survey respondents included automated red flags or business rules (64%), scoring capability (60%) and link analysis (57%). In most cases, these tools automate many manual tasks – business rules, for example – associated with fraud detection.
“However, less than half are using technology for application/underwriting fraud or internal fraud,” the report found. This would include detecting fraud through risk assessment and/or point of sale.
Also, insurers are generally not using more sophisticated technologies that are available. Fewer than half of the insurers surveyed, for example, use workflow routing (43%), text mining (40%) predictive modeling (40%) or geographic data mapping (23%).
Insurers indicated they were most likely to invest in predictive modeling (33%) and text mining (31%) over the next 12 to 24 months, the study found.
SAS and the coalition noted that the anti-fraud technology still relies heavily on the quality of the insurer’s data, with the most common sources of data being the carrier’s own claims data (69%) and public records (62%).
Limitations to such data include the fact that information silos are still prevalent in the insurance industry, and many organizations are still using a combination of legacy systems, spreadsheets and internal databases. Partly as a result, many insurers still haven’t been able to fully integrate anti-fraud technology with claims and other systems.