In many ways, the early 1990s were the heyday of Special Investigation Units (SIU) for insurance companies, says Tim Wolfe, Director of Special Investigations for CNA, a commercial property and casualty insurance provider. Around that time, states started requiring companies to report suspicious claims and insurers increased staff to meet those regulations. “In those days, SIU consisted almost entirely of former law enforcers. We were hiring people who were used to policing crime.”
Today, analytics technologies that identify fraudulent activity are changing all of that. Insurance companies are hiring data modelers instead of security professionals, and they’re changing the way SIU departments are organized.
“That doesn’t mean you don’t need boots on the street,” says Wolfe. “You still need field investigators asking the right questions.” Technology does not do investigations for you, but technology will identify potential fraud activity that may have been overlooked so you know where to send those investigators and waste less of their time.
The staffing change is reflected at CNA dramatically. In fact, Wolfe’s team has outsourced a large portion of its field investigative work and now employs a handful of workers who manage and monitor the investigative process, plus an in-house team of investigators for major investigations and organized crime. The rest of his team is focused on intake, training, regulatory compliance, analytics and process improvement.
“The last data analyst we hired was from the military,” says Wolfe. “Her experience was in predicting IED explosions, but we’ve found that it was a really good idea to hire somebody outside of the industry. They’re not just thinking insurance and they bring a lot of fresh ideas.”
The change in skill sets is reflected at a national auto insurance company in the U.S., where a senior manager of the Special Investigations Unit tells us they have transformed the home office environment over the last couple of years to be an innovation environment. As a result, his team is really starting to look at how data can influence decisions.
Using data to identify fraud
Data has been used traditionally in underwriting and claims side of most insurance businesses, but it’s a recent practice to use data to manage resources and flag potential fraudulent claims for investigation. There are clear, distinctive data points that can be used for modeling. If businesses can identify these data points and convert them into operational decisions while collecting information from customers, they can be more predictive.
In the past, most fraud referrals came from a small percentage of adjusters who know what to look for, take the time to make reports, or have good instincts. Today’s fraud systems, however, can identify red flags automatically to help make adjusters and SIU unit aware of potential fraud more often and more quickly.
Another fortunate turn in identifying fraud in insurance is that regulators are now allowing insurance companies to share info as it relates to fraud. For example, a Medical Crimes Database is being built with impetus from the National Insurance Crime Bureau in the U.S. where all major insurance companies can share medical code records (in aggregate without revealing individual patient info) to better detect fraudsters who are making claims across multiple insurance companies.
The value becomes obvious when you see a single person claiming 60 hours of work in a 24 hour day. If you’re only looking from one company, you might see five hours of it – but looking across every claim, the crime becomes obvious.
Wolfe says technology is also important because fraud is getting more sophisticated than it used to be. “We’re seeing more granulated schemes involving not just doctors and lawyers but people associated with workers compensation who bill for interpretation services when they’re not needed. Or transportation providers to drive patients for care when no transportation has took place.”
Ultimately, identifying fraud helps everyone. It reduces business costs, which can be passed on to the customer and it helps fight larger societal criminal elements. Plus, claims from legitimate customers get processed more quickly and fewer false positives keep honest customers happier too.
