SOCIAL MEDIA & SURVEILLANCE CAMS: NEW TOOLS FOR DISCOVERY

"There needs to be more unstructured data analysis in fraud analysis—and by that, I mean claims-adjuster notes, data from the Web, data from social media," says Kimberly Harris-Ferrante, vice president and distinguished analyst at IT research and advisory company Gartner.

Social media has been a hot topic within SIUs in recent months. With nearly two-thirds of adults maintaining profiles on social-networking sites, mining social data for evidence of activity that contradicts a statement of injury or loss can be invaluable in preventing fraudulent claims from being paid.

"People are very quick to post all kinds of information and say some things [in social media] that they probably shouldn't have," says Mike Butler, commercial market SIU manager for Liberty Mutual.

About a year ago, Secura Insurance hired a dedicated social-media investigator. In one case, the insurer found that an individual claiming he could no longer golf had posted publicly on Facebook about being a member of the latest championship team at his country club. In another, a person claiming total disability posted pictures of himself riding a skateboard. In both instances, social-media data helped Secura resolve the case.

The proliferation of surveillance technology today has also been a boon to SIUs.

"There are cameras everywhere. They aren't the silver bullet, but they do provide evidence," says Butler. In a recent case, Liberty Mutual was able to utilize traffic-camera images to prove that a person claiming to be injured in an accident was not in the vehicle. 

OLD MODEL, FRESH APPLICATION

Despite the role of a surveillance camera in the aformentioned case, Liberty Mutual still gets more bang for its SIU buck from more traditional data-crunching and workflow technologies.

Like many carriers, it has a red-flag system to bring suspicious claims to adjusters' attention; the difference is that Liberty Mutual uses an internally built fraud model to set flags.

The company is understandably circumspect in how much detail it will reveal about its propriety model, but it reports that it leverages its extensive book of Workers' Comp business to provide relevant data.

"Using [claim attributes] that are established red flags in the industry is just the start," says Steve Friedman, national director of special investigations for Liberty Mutual Agency Corp. "We use predictive modeling to create a model that is forever adapting, from first notice all the way through the life of the claim."

Gartner's Harris-Ferrante confirms that in today's environment, creating an effective fraud model is not a "set-it-and-forget-it" activity.

"Fraudsters get smarter over time, so you need a cycle where fraud is identified and confirmed, and it becomes a learning process. You need to be open to identifying new models—understanding that there might be better triggers that are determined through data analysis or investigation of outliers," she says.

At State Farm, the evolution and adoption of link-analysis tools, which detect trends across hundreds or thousands of claims, has increased the company's ability to combat organized fraud.

"Previously, a fraud-ring investigation with 100 claims would consume a lot of time to gather information needed to make decisions. Now, we can do a 500-claim investigation in a matter of hours," reports Schulkins.

Although State Farm won't disclose the specific third-party analysis platforms it has chosen, Schulkins says these tools have become essential.

"[Fraudsters] have become more sophisticated in hiding their schemes," he explains. "The tools will not make a claim decision for you, but they give the decision-maker the most relevant information quickly."

FIGHTING ON MULTIPLE FRONTS

For its part, Nationwide believes that fraud-fighting is closely tied to reducing overall theft losses, which is why both activities fall under the company's SIU.

"We're here to protect the consumer, and it doesn't matter if [those consumers are] our customers or the general public," says Lynne Brady, associate vice president of special investigations at Nationwide.

One of Nationwide's more successful initiatives is its bait-vehicle program, consisting of more than 70 vehicles in 19 different states that are hot among thieves.

Nationwide analyzes its own loss data to determine where there is significant theft, then partners with different local, state and national agencies to supply bait cars. Technology allows vehicles to be driven a short distance by thieves before they are automatically locked down, with their location transmitted via GPS to law enforcement.

"There have been over 400 arrests [due to the program], and in certain areas we've seen theft go down anywhere from 20-60 percent," says Brady.

In addition to the bait-vehicle deterrent program, Nationwide uses post-claim technology in its SIU and throughout its claims operations, including license-plate reader (LPR) cameras.

"We have been able to recover hundreds of vehicles, not just Nationwide's [insured] vehicles" in the LPR program, Brady adds.

Although designed primarily to reduce theft loss, programs such as these enhance a company's reputation of being "tough on crime," which can positively impact fraud-fighting efforts. Data supports this strategy: More than two-thirds of consumers say they believe insurance fraud happens because people believe they can get away with it, according to statistics cited by CAIF.

"You need to build a company culture where it's known that fraud is not tolerated. You get that reputation by bringing affirmative legal actions against certain types of fraudulent claimants," says Kathleen Gordon, national coordinating counsel of Liberty Mutual's SIU. 

Ultimately, investigators understand that technology is just a weapon in their fraud-fighting  arsenals. "All of these tools provide information, trends, behavioral comparisons and so on," says Schulkins. "But in the end, those are just leads to point you in the direction of where you need to deploy human manpower."