Fraud costs the insurance industry $80 billion a year, but insurers now have a powerful weapon to combat it.
A conservative estimate from the Coalition Against Insurance Fraud pegs the annual cost of fraudulent insurance claims at $80 billion across all lines, with $34 billion of that amount coming in the property/casualty sector. The Coalition further reports that 5 to 10% of claim payments by U.S. and Canadian insurers were from fraudulent submissions.
Statistics from the FBI further highlight the massive cost of insurance fraud. The agency that estimates non-health-insurance-related fraud amounts to more than $40 billion per year, adding between $400 and $700 each year to the average U.S. family’s premiums.
With so much money at stake, insurers have more incentive than ever to eliminate and prevent insurance fraud. That’s why they’ve turned to data analytics programs to proactively identify insurance scams as they unfold.
Using Technology to Investigate Fraud
In 2016, the Coalition Against Insurance Fraud and analytics firm SAS polled 86 market-leading property and casualty insurers on their use of technology to detect fraud, finding that 75% of respondents had set up automated systems to uncover false claims. That percentage was a significant increase from surveys done in 2014 and 2012.
The Coalition survey further revealed that roughly 90% of insurers employ business rules and red flags to automatically detect suspicious claims, and that the use of predictive modeling and link analysis has climbed 21% over the past two years.
Another survey of U.S. property and casualty insurance companies by WNS DecisionPoint, a thought leadership and research platform serving various industries, found that 73% of respondents had invested in data analytics, yet only 38% of them were using it to its fullest capacity. Those insurers that had employed data analytics to catch fraud lowered their investigation costs by an average of 1.4 times compared to insurers that did not.
WNS DecisionPoint details the four stages at which insurers can detect fraud using data analytics: point of sale, first notice of loss (FNOL), investigation, and post claims. At each stage, a data analytics program can scrutinize the submitted claim for any indications of fraudulent activity. Basic anti-fraud solutions, for example, immediately certify an applicant’s identity and whether the person has previously been involved in fraud.
During the FNOL stage, a data analytics program can comb through a large number of claims to spot any fraudulent activity while validating legitimate submissions. Later in the process, it can greatly shorten investigations by quickly linking claimants to suspicious behavior. After claims have been paid, an analysis of the data collected enables insurers to spot any organized fraudulent schemes, helping them identify fraud in the future.
Collecting the Data
Detecting insurance fraud, of course, depends on collecting the right type of data, and that entails weaving together both internal and external information. Customer information and policy provisions would be classified as internal data held by an insurers, but insurers can also capture external data such as weather conditions at the time of a claim event.
Such external data may include information gathered from Internet of Things (IoT) devices, health sensors, and other smart technology. A home equipped with alarm system would confirm whether a burglary occured at the property — or if the owner is attempting to file a false claim. A fitness tracker corroborates if a claimant was at the scene of an accident or fire. With this information gleaned from IoT and run through data analytics software, insurers now possess an even more powerful weapon in the battle against fraud.