The insurance industry has focused the majority of technologyinvestment for automated fraud detection efforts around autoclaims.

This makes sense: Auto coverage represents 38 percent of totalproperty and casualty exposure, or $200 billion in direct writtenpremiums, according to the SNL Briefing Book. However,Homeowners comprise the second-largest exposure at $93 billion indirect written premiums. Fraudulent losses involving fire, theft,water, scheduled property, wind or hail are exposures in which lossseverity can be significant and greatly affect the bottom line.

Here's a prime example: What if there were an opportunity toturn a $300 investment into $30,000 in 60 days or less with minimalrisks? Certainly, this ROI is far better than anything conventional(or unconventional) investments can offer.

In this case, the investment is a premium for a scheduledproperty insurance policy jewelry rider. The return is a fraudulentclaim payout for insured jewelry items that may have never evenexisted. It is easy, and the penalties for the few who get caughtare minimal.

Unfortunately for the property insurance industry, thismentality and perceived opportunity by some policyholders creates avery real challenge. However, claims volume, customer retentiongoals, growing centralization of claims handling operations, thelegal environment and service-level metrics all make it challengingto effectively detect and investigate suspicious propertyclaims.

What can the industry do?

One key foundational strategy is to pursue automated frauddetection technology.

To date, automated fraud detection options for homeowners'insurance are limited. However, a lot of experience, knowledge andlessons learned from advanced analytic detection efforts in autoinsurance can be applied to make homeowners coverage a lessattractive "investment" for fraudsters.

One of the first lessons is to use an organization's mostvaluable asset — data. Carriers invest an incredibleamount of resources into gathering data through direct customerinteraction, third-party data and industry information. In fact,the SNL Briefing Books says that approximately $23 billion wasspent on underwriting expenses and $9 billion adjusting lossexpense in 2015, or 34 percent of $93 billion in direct writtenpremium for homeowners. 

Advanced analytics represents the best opportunity to capitalizeon this investment. The challenge is that this disparate data issiloed across the organization and difficult to access. Technologymust be able to ingest, cleanse, enrich and resolve discrepancies.Once this enterprise-wide view of data is accomplished, advancedanalytics can be applied.

Related: 8 ways social networks help identifyfraud

Fraudulent homeowners' claims usually occur within theinventory claim (Photo: iStock)

For example, by using public records information such asbankruptcy, liens and judgments combined with loss history andapplication information, you can derive some extremely powerfuldetection metrics and anomalies. Applying this to the entirepopulation of claims from the first notice of loss can surfacethose questionable property loss cases much earlier in the claimslife cycle for appropriate inquiries and deeper investigation. Thisis especially true for the jewelry loss example. Another lesson isdays to detection for high scoring: Egregious losses can decreasefrom an average of 42 days down to one day. This allows for earlyintervention and improved triage.

Effective user interface

Insurers can gain further efficiencies through an effective userinterface. Aggregating insightful, clear information into one placehas allowed for quicker triage, appropriate assignment and focusedinvestigation planning. Link charts have proven effective inidentifying complex or previously unknown relationships that canallow for better focus and deployment of investigativeresources.

The usefulness of this capability in the homeowners space hasgreat implication for identifying organized networks ofcontractors, insureds and other vendors. This would be particularlyuseful after a catastrophic event.

Finally, it is important to score a claim from first notice ofloss through final disposition. Homeowners' losses, especiallythose involving personal property inventories of claimed items,develop over time. It may be many months before documentation issubmitted after a first notice of loss. For theft, the fraudgenerally occurs within the inventory claim. Having a means toscore a claim as this insured documentation is submitted providesanother critical detection capability.

Homeowners' insurers have a green field of opportunities toimprove detection processes. Insurers that have implementedadvanced technologies have seen fraud mitigation improve by 25percent or more. It will take creativity and the right investments,but ultimately, making your premium a less attractive investmentfor fraudster is paramount.

Related: Walking a fine line with soft frauddetection

Jim Hulett is an insurance solutions consultant in theGlobal Fraud and Financial Crimes practice at SAS. He may bereached at [email protected].