The state of insurance fraud technology

Internal fraud, rate evasion, underwriting fraud, claims fraud, cybersecurity fraud … the various flavors of insurance fraud add up to more than $80 billion a year in the US, according to conservative estimates from the Coalition Against Insurance Fraud¹. Fraud comprises about 10% of property-casualty insurance losses and loss adjustment expenses – about $34 billion a year².

In 2018, the Coalition Against Insurance Fraud, in partnership with SAS, conducted its fourth semiannual study to better understand how insurers are using technology to detect and prevent these losses. The research drew on online surveys, mostly with property/casualty insurers, and interviews with senior insurance executives and other industry experts. To what degree are insurers succeeding with anti-fraud technology? What are their challenges, results and future plans?

The 84 participating companies represented a significant share of the property/casualty market. Because this study included more life, health and disability insurance companies than in previous years, some 2018 calculations were weighted to enable more accurate comparisons over time.

Here are some key findings from the March 2019 report, The State of Insurance Fraud Technology:

Fraud is up

For the third consecutive study period, insurers report higher levels of suspected fraud. Nearly three-quarters of respondents (up 11 percentage points since 2014) say fraud has increased either a little or a lot. No insurer in the last six years has reported that fraud has decreased significantly.

More insurers have adopted anti-fraud technologies

As more tools become available and costs decline in some areas, more insurers have expanded their anti-fraud arsenal of technologies and integrated them into their underwriting systems. The growing range of outsourced services is a particular advantage for smaller companies, which have more limited budgets and IT resources.

Further investment is ahead

Two-thirds of respondents (66%) plan to acquire new claims fraud detection technology in the next year, while 34% plan to beef up their tech defense for underwriting fraud, and nearly a quarter plan to enhance their ability to deal with cyber fraud.

Anti-fraud technology budgets are up

More than 40 percent of insurers have bigger budgets for anti-fraud technology in 2019 – the highest percentage since these studies have been conducted. Predictive modeling and link analysis/social network analysis are the two top areas where new money will be spent. Only 2% reported a shrinking tech budget in 2019.

Claims fraud is still the chief focus

Every property/casualty company (and half of non-PC insurers surveyed) say detecting claims fraud is their primary reason for using technology – a significant increase from 2016. Even more dramatic, 55% of insurers are using technology to uncover underwriting fraud, a figure that has doubled in the last six years.

The defense is more sophisticated

As use of anti-fraud technology matures, most insurers have shifted from single technologies to blending tools to gain new capabilities and refine detection techniques. As they opt to use newer analytics, many are relying much less on old-school technologies such as automated business rules and red flags.

Advanced analytics is taking center stage

The biggest areas of expansion for the next two years are predictive modeling (64%), link analysis/social network analysis (43%), text mining (36%) and exception reporting/anomaly detection (32%), while 21% of insurers, including half of non-PC insurers, said they plan to invest in artificial intelligence in the next 24 months.

Unlike rules-based systems, which are fairly easy for fraudsters to test and circumvent, machine learning – a form of AI – adapts to changing behaviors in a population through automated model building. With every iteration, the algorithms get smarter and more accurately detect fraud.

Lots of data remains untapped

For the most part, there was little change from 2016 to 2018 in the data sources insurers use for fraud detection and investigation. Property/casualty insurers are most likely to use third-party data (80%). Notably, only 61% of insurers use industry fraud alerts and watchlist data, down from previous years.

Data integration remains an issue

More than three-quarters of respondents (76%) struggle with integrating data into their systems, up from 64% in 2016. Data integration challenges likely stem from the many different sources and formats of data now available to insurers, as well as the shortage of qualified staff adjusters who understand how to integrate data.

IT resources are a bottleneck

Insurers have been citing IT constraints as a major issue since the first study in 2012. As outsourced options become more plentiful and affordable, more insurers are looking outside the company. In the 2018 study, only 47% of anti-fraud systems were built and maintained in-house, down from 64% in 2016. This trend is likely to continue.

Referrals from technology have plateaued

This finding was a surprise. About 20% of insurers report receiving 30% or more of their referrals from technology. A possible reason is that insurers have tweaked their systems to generate better-quality referrals at the cost of generating more referrals that, in some cases, their SIUs cannot handle due to the large volume.

Those referrals are better-quality

When asked about the benefits they get from anti-fraud technology, 68% of respondents cited the quality of referrals. Higher-quality referrals mean fewer false positives and false negatives, which lead to the next-most commonly cited benefits: mitigation of losses and more effective use of investigators’ time.