Insurers Using Technology to Fight Insurance FraudJanuary 5, 2017 |
Insurance fraud has been estimated by some to be as large as an $80-billion-per-year problem. See, e.g., Coalition Against Insurance Fraud, “By the numbers: fraud statistics.”1 To fight insurance fraud, insurance companies, working hand-in-hand with their lawyers, have begun to adopt new kinds of technology. These tools—ranging from social media to data analytics—are being used by insurance companies to help stop fraudulent policies from being issued or renewed based on false representations of fact and to block false, inflated, or otherwise fraudulent claims from being paid.
A substantial portion of insurance companies now are including anti-fraud technology in their anti-fraud programs. Indeed, the Coalition Against Insurance Fraud recently surveyed insurance companies and found that nearly 75 percent have fully integrated technology into their anti-fraud systems—up from about 50 percent four years ago. See, “The State of Insurance Fraud Technology.”2
Imagine a person filing for disability benefits and, two days later, participating in a bike race. Then imagine that, after the race is over, race organizers post the person’s standings online.
Think of a woman who collected on an insurance claim for her lost wedding ring and later is seen wearing the very ring in a photo available for all to see on the Internet.
Or consider an Internet search revealing a semi-professional athlete playing for his team after he had claimed that he had been injured, could not work, and was entitled to Workers’ Compensation benefits.
These examples, highlighted in a recent white paper by insurance company QBE North America titled, “Innovations in Using Social Media to Fight Insurance Fraud, Improve Service,”3 illustrate the power of social media to thwart fraudulent insurance claims.
As QBE observed in its white paper, most adults use social media today, leaving a public trail of what they have said and done, and with whom, that may conflict with what they have told their insurance companies when filing a claim for benefits under an insurance policy. Insurers have taken notice, and many now have expanded their special investigation units (SIUs) to include social media analysts who know how to search and make sense of information in databases and on public websites and social media platforms. QBE emphasized in its white paper that these analysts are experienced claims or SIU professionals and “not hackers,” and that they work within well-prescribed legal and procedural guidelines.
Social media analysts working at insurance companies can help to determine whether an accident happened as an insured said it did, whether an injured person who claimed to be unable to participate in one or more daily activities was in reality fully able to do so, and what the claimant’s social media connections suggest as to his or her claim. Simply put, as Dan Franzetti, the chief claims officer for QBE North America, said in a statement, “[L]everaging the power of social media helps us more effectively and precisely identify instances of fraud to drive the cost out of the system.”
Data Analytics, Automation
As a practical matter, a social media analyst working alone, or even in a large group, would not be able to review the sheer volume of information that is posted on social media platforms every day. It would be overwhelming. To help, insurers have been turning to data analytics and automation to detect fraud. As the Coalition Against Insurance Fraud reported in its paper, “The State of Insurance Fraud Technology,” the most popular anti-fraud technology used by insurers today is automated red flags/business rules, which connect with existing claims procedures to isolate and highlight suspect claims and which is used by 90 percent of insurers that use anti-fraud technology.
Other kinds of this technology currently in use by insurers include predictive modeling, which is used by more than half of insurers and which involves the creation, testing, and validation of a model to predict the probability of claims being filed; exception reporting, which identifies claims outside the range of what is considered normal; text mining, which involves the process of devising patterns and trends to derive high-quality information from text; geographic data mapping, which helps to determine patterns or trends outside the expected range in a geographic area; and data visualization, which puts data into a visual context to help it be more easily understood.
Insurers have a wealth of data available for use with their fraud detection systems. According to the coalition, internal data and public records are the largest sources of information used by insurers. Insurance companies also use third-party data and data aggregators as well as social media data. The coalition also found that integrating industry fraud alerts and watchlists into insurer anti-fraud systems is much more common now than even two years ago.
Insurers surveyed by the Coalition Against Insurance Fraud for its paper, “The State of Insurance Fraud Technology,” said that they believe that anti-fraud technology has improved their efforts to fight insurance fraud. Insurance companies told the coalition that technology has produced more referrals for possible fraud, and better-quality referrals. Seventy percent of insurance companies surveyed said that technology now accounts for more than 10 percent of fraud referrals, and 6 percent said they receive more than 60 percent of their referrals through technology.
Insurers also told the coalition that they measure the value of their anti-fraud technology in a number of ways. Primarily (50 percent), they look to the fraud detection rate as evidence of the technology’s worth. They also have found the number of fraud referrals to be telling, as well as the number of days from the first notice of a claim to detection of possible fraud.
Issues to Address
Insurance companies surveyed by the Coalition Against Insurance Fraud for its paper on “The State of Insurance Fraud Technology” acknowledged that they must address a number of issues relating to anti-fraud technology.
First, they noted that there are limits to the amount that they can spend for anti-fraud technology—and for in-house talent to use the technology—especially given the other uses that insurance companies have for technology and the growing demand for technology experts at insurance companies and elsewhere. Nevertheless, one-third of insurers are expecting anti-fraud technology budgets to increase in 2017, with half saying it will remain flat and the balance indicating a decrease. The anti-fraud tech funds, insurers told the coalition, will be spent on predictive modeling; social media software; text data mining; and link analysis, which is a technique used to evaluate the relationships or connections between policyholders and claims.
Second, insurers expressed some concern to the coalition over “[e[xcessive false positives.” The coalition reported that SIU directors said that their units spend “too much time investigating cases that are not legitimate fraud reports.” However, it added, as insurers gain more experience with tools such as automated red flags and predictive modeling, they are likely to generate fewer false positives.
Next, insurers are concerned about their ability to integrate data into their systems and poor data quality. One can expect that, with the growing use of technology, this, too, will likely become less of a concern.
Of course, insurance companies also are sensitive to consumers’ privacy concerns raised by the use of anti-fraud technology and have to ensure that their own data is safe from hackers and other potential security problems.
Anti-fraud technology is playing an important—and growing—role at insurance companies. In addition to the tools already discussed, one can expect to see more use by insurance companies of telematics, such as Internet-enabled automobiles that can help to determine the cause, location, timing, and severity of automobile accidents. So-called “wearables,” which track users’ activities, can transmit data to insurers in real time and may grow in use for insurance purposes in the future. And, even now, drones are helping insurance companies to evaluate property and casualty claims and to reduce fraud.
These tools, when combined with appropriate staffing and training, may slow the growth of insurance fraud, resulting in lower premiums for consumers.
1. Available at http://www.insurancefraud.org/statistics.htm#.V1HXcE1wW71.
Reprinted with permission from the January 5, 2017 issue of the New York Law Journal.
- Evan H. Krinick