Insurance fraud remains a large drain on insurers, and on consumers who bear high personal and financial costs this crime imposes.
This is why the Coalition surveyed 86 insurance companies about how they deploy anti-fraud technology. Combined, they represent a significant share of the property-casualty market.
The outcome is the newest review of how insurers use anti-fraud analytics, “The State of Insurance Fraud Technology.” We sought a data-driven profile of tech usage — tools, how insurers use them, perceived benefits and obstacles to further adoption.
A majority of insurers agreed that fraud has increased against their company over the last three years. More insurers also embrace and have expanded their use of tech systems as a key building block of their anti-fraud strategies.
The Coalition conducted similar studies in 2014 and 2012. The current study compares how insurance fraud has changed, and how advances in technology enable insurers to better combat insurance crime.
Nearly 75 percent of insurers have fully integrated technology into their anti-fraud systems, the current study finds. Clearly, insurers are more-comfortable using technology and justifying its expense. Insurer senior management is more analytically astute. Leadership increasingly believes investing in technology helps improve their company’s competitive advantage.
The belief in higher fraud levels may be a big reason insurers justify a larger investment in anti-fraud technology. Some 61 percent say suspected frauds increased slightly or significantly in the last three years. This compares to 51 percent in 2014.
A related question is whether fraud is increasing or insurers are getting better at detecting this crime — partly through greater use of technology.
Among other findings:
• Detecting claims fraud continues as the leading area for technology usage. Some 76 percent of insurers use technology to uncover suspect claims jumped to 76 percent;
• Automated red flags/business rules are the most- popular tech tool. Some 90 percent of insurers that use anti-fraud technology also use red flags/rules;
• More than half of insurers surveyed use predictive modeling, a significant increase over just two years ago;
• Internal data and public records remain the largest information sources for populating technology systems;
• Technology is uncovering more referrals, and better-quality ones. Increased loss mitigation is another oft-cited benefit;
• 70 percent of insurers said technology accounts for more than 10 percent of their fraud referrals;
• The two biggest challenges insurers face: lack of IT resources to maintain and expand programs, and excessive false positives; and
• One-third of SIU directors expect their IT budgets to increase in 2017. Tops on their shopping lists are predictive modeling and link analysis/social-media programs.
State of fraud & tech
The full scale of insurance fraud is unknown. This crime often goes undetected, so the anti-fraud community can only estimate the volume of this crime and dollar losses.
“Tops on their shopping lists are predictive modeling and link analysis/social-media programs.”
Fraud also exists throughout the insurance lifecycle, from the application process through claims resolution. Insurers increasingly see more attempted fraud at “point of sale” — during the application and renewal. This is most-common with online coverage purchases. Insurers also combat internal fraud, money laundering and, for the last few years, the emerging issue of cyber fraud.
Areas employing tech. Some 76 percent of insurers said detecting claims fraud is the primary use of anti-fraud technology. That compares to 65 percent in 2012. Using technology against underwriting and automobile rate-evasion schemes is increasing as well.
Using tech to uncover internal fraud has plateaued at 29 percent. Insurers using anti-money-laundering software fell from 24 percent to nine percent over the last two years.
Cyber fraud continues to be a growing concern for insurer anti-fraud departments. Nearly one of five say they use technology to combat this threat.
Tools employed. Most insurers agree that a combination of tech tools usually is required to identify schemes — whether committed by complex crime rings or smaller-dollar opportunists.
The first line of defense continues to be automated red flags/business rules. They are the bread and butter of anti-fraud technology. They can quickly help insurers tag honest claims for payment, and isolate suspect ones for a closer look. In fact, 90 percent of insurers surveyed reported using automated red flags and business rules, up from 64 percent in 2012.
The use of predictive modeling also increased significantly as more insurers went online with this technology. The percentage of insurers using predictive modeling jumped from 40 to 54 since 2012.
Link analysis and mining social media also saw substantial increases. Two-thirds of insurers surveyed said they use these tools.
Usage remained largely flat for exception reporting or anomaly detection, text mining, geodata mapping, data visualization and case-management systems. While the number of users has risen likely because of the larger sample size in 2016, the percentage remained the same. The 2016 study also includes a larger percentage of insurers that are later adopters of anti-fraud technology — another reason for the potential lag in apparent growth of tech usage.
Insurers also were asked how often they refresh their automated red flags/business rules. The most common answer was annually (34 percent), though 32 percent said they refresh more often.
“Integrating industry fraud alerts into internal systems also is becoming much more prominent …”
Sources of data. Insurers report plenty of options to feed data into their systems. Data sources have expanded with more data vendors coming online, and as insurers find greater use of internal data from claims systems and elsewhere in their own files.
While all data sources have increased from prior studies, internal data, public records and social media have grown the most. Integrating industry fraud alerts into internal systems also is becoming much more prominent since 2012 and 2014. The increased availability of alerts likely is encouraging more insurers to integrate them into their systems.
Workflows will become more proactive as the quality, quantity and variety of data expand — in conjunction with the ability to automatically scrub data. That will enable SIUs to focus on the most-significant threats.
Benefits of anti-fraud tech
Most insurers said receiving more referrals, better referrals and increased loss mitigation are the top three benefits of their tech systems. Those findings are similar to 2014. Uncovering complex or organized rings and improving investigator efficiency were cited less often than previously.
Fully 70 percent of insurers said they receive more than 10 percent of referrals from technology — up from 66 percent in 2014 and 55 percent in 2012. Interestingly, no insurers reported receiving more than 60 percent of referrals from their automated systems in the two previous studies. Six percent of insurers reported so in 2016.
Insurers often ask if there is an optimum range of referral percentages their systems should produce. There is currently no standard optimum, most experts agree. Results largely will continue to depend on sophistication of systems, training of users, claims philosophy and mix of business.
However, a meaningful benchmarking metric might be developed when anti-fraud tech matures and is used more-uniformly.
Challenges of implementing tech
The top challenges of employing technology are similar to 2014:
• Limited IT resources — budgets and in-house expertise — ranks tops. Technology is expanding rapidly in most areas of insurer operations, from marketing to underwriting to legal. The demand for internal IT services is high. Yet budgets for outside services are too low for many companies both to maintain existing tools and add new ones
• Excessive false positives are the second-most-cited challenge. SIU directors say their units spend far too much time investigating cases that turn out to be red herrings. While insurers vet most leads during triage, excessive false positives waste valuable resources that are in short supply in many SIUs.
Excessive false positives are more likely to be a problem for insurers that use a narrow range of technologies and/or data. Insurers using a robust range seem to experience fewer false positives.
There also is growing anecdotal evidence that the more experience insurers gain with their systems — especially with automated red flags/business rules and predictive modeling — the more they can reduce false positives by tweaking the systems. Insurers talk about reaching a “sweet spot” where their systems uncover a high level of suspect claims while generating fewer false positives.
“Excessive false positives are more likely to be a problem for insurers that use a narrow range of technologies and/or data.”
Justifying the benefits of using anti-fraud technology appears to be less of a problem for many insurers. It was the biggest challenge cited in 2012. Insurers are growing more comfortable with technology. Thus it appears SIU leadership and senior management understand the positive business benefits of using technology against fraud. The growing consensus: tech uncovers more crime, faster and with greater accuracy.
Fraud-detection rate was the most-cited metric for measuring success with technology, and next came the number of referrals. Interestingly, one of five insurers said they do not use metrics to gauge tech success.
Another potential measurement includes number of days from first notice of claim to detection. Automating detection to assess all claims from first notice allows early claim-cycle detection opportunities to reduce losses.
Anti-fraud technology likely will continue growing in usage through next year. Nearly a third of insurers say they are budgeting to expand their technology. In 2014, only a quarter of insurers said they expect bigger budgets for the next year, so it appears technology investments are accelerating.
Most insurers say they will invest in predictive modeling, followed by link analysis and social-media software, and then text mining.
Today’s anti-fraud technology continues to expand and become more-effective. Just as important, it evolves with shifting fraud schemes. Software has advanced to where it can “learn” from experience, getting better at detecting fraud and identifying patterns. This “learning” enables software to continually adapt and increase in sophistication as it gathers more data. The smarter the tools, the greater chance of detecting fraud early on, and even of predicting scam activity before criminals uncover the opportunity themselves.
Many insurers use the buzz phrase “speed of detection.” This describes an aspect of technology that convinces more claims handlers to embrace these new tools. Suspected frauds take extra time and work for claims staff, and lengthen cycle time. A natural tension exists in many insurance companies between claims departments that focus intently on closing files, and SIUs that want to slow the process and investigate.
Newer technologies such as predictive modeling can meet both goals. They help detect fraud earlier in the claim process — thus shortening cycle time. Conversely, tech more quickly validates legitimate claims and allows insurers to pay them more promptly.
While referrals from claims staff always will factor into anti-fraud workflow, current and future technologies likely will accelerate fraud detection, thus allowing faster resolutions of legitimate and suspect claims.
Discussions with insurers that derive excellent results from anti-fraud tool underscore the importance of having knowledgeable and well-trained staff to use and support tech tools to their fullest degree. Results will fall short of expectations unless tech is used in conjunction with investigators’ instincts and savvy.
Insurers that embrace the right mix of tools, staffing, training and technologies will continue seeing reduced claims costs, more-accurate pricing, a competitive edge and lower premiums for policyholders.
About the study
The “State of Insurance Fraud Technology” was undertaken by the Coalition to better understand how and how much insurance companies use anti-fraud technology. It addresses anti-fraud technologies insurers now use, and are considering using. The study follows up similar research conducted in 2012 and 2014.
Technical assistance was provided by SAS Institute, an international company focusing on technology solutions for businesses and governments.
Technical review and oversight was provided by the Coalition’s Research Committee: John Kloc, Sentry Insurance … David Rioux, Erie Insurance … Steve Friedman, Liberty Mutual … Jack Dever, American Family … Joseph Theobald, Citizens Property Insurance Corporation.
The research drew on two initiatives:
• Online survey with 86 mostly property-casualty insurers providing data in June and July 2016; and
• Qualitative research, including in-depth interviews with a range of subject-matter experts and senior insurance executives.
The Coalition thanks all who cooperated, for their time and insight.
About the author: Dennis Jay is executive director of the Coalition Against Insurance Fraud.