Consumers want a friction-free quote and application process, and insurers have delivered on that expectation across the globe, standing up online quote and application platforms.
The speed of digital insurance experience comes at a price: Almost 70% of insurers agree that increased digital activity leads to increased fraud. While data integrations play an important role in spotting some misrepresentation and fraud during applications, an estimated $35B remains undetected
AI solutions can address the specific challenge of finding hidden fraud and misrepresentation without adding friction to the application process. To do this, these solutions need to apply different techniques throughout the policy lifecycle.
Fully vetting new policies – without slowing down applications
Insurers experience a few risks risks regarding new applications:
Using AI, it becomes possible to solve all these risks, checking for the gross characteristics of fraud without interrupting the customer experience.
AI solutions can check an application against several first-party or external databases in real time, identifying red flags such as prior fraud, fraud network connections, or bad actors attempting to pass off vehicles with salvage titles as being in good condition.
With misrepresentation, AI solutions can mitigate risks by checking an application against known facts from the same sources. As an example, think of an applicant that tries to get a lower rate on home insurance by misrepresenting facts about their property. Here, the solution can pull data from municipal records – to find renovation permits and satellite imagery – to assess the condition of the building at the time when the policy is being written.
AI solutions aren’t just smart – they’re built on scalable cloud platforms with robust processing capabilities. What this means is that application risk checks take place in real time, clearing honest applicants for accelerated policy binding..
Detecting underwriting fraud while a policy is in force
While AI solutions make it easier to detect individual fraud during the application phase, fraud networks and other more sophisticated bad actors require even more sophisticated analysis. That’s because organized fraudsters will make more of an effort to conceal their activities.
After a policy is sold, underwriters have up to 90 days (depending on state laws and operational procedures) to review new policies for fraud. This should give them enough time to find and investigate organized fraud, but AI solutions make this process much easier.
At this stage, AI solutions can solve problems by deploying more complex investigation scenarios. Since analyses don’t have to run in milliseconds, AI solutions can run in batches. Typically, the AI will ingest new policies collected in a 24-hour period, run 24 hours of analysis, and then flag suspicious accounts.
In an agent gaming scenario, for example, an agent may apply favorable ratings, mileage, or discounts across many of their policyholders to provide a lower premium. This increases their chance of a sale. Viewed as a single application, there may be no indicators of misrepresentation. AI analysis can uncover this anomalous pattern by comparing agent peers and comparable policyholders. It can then highlight the discrepancy for investigators.
Renewal period – detect undisclosed policy changes
Lastly, best practice is for insurers to check on policies at point of renewal or adjustments. That’s because some policyholders forget (or “forget”) to update their information based on life changes. Circumstances may change in a way that would cause a premium to increase, but there’s no guarantee that the policyholder will report this on their own.
As one example, let’s say that a policyholder moves to a more urban location where there’s a greater chance of a car accident or theft. The policyholder may decide not to report this information, which would allow them to continue paying a lower premium.
Fortunately, there’s a lot of information that can flag the move. The AI solution can access enough of this information to make the inference, and then flag an underwriter so they can ask the question, “did you move in the last twelve months?”
Shift Underwriting Risk Detection adapts to any stage of the underwriting process
Underwriters face different challenges to risk detection throughout the policy life cycle. It’s difficult to review policies under strict deadlines, but Shift’s AI solution can analyze, triage, and report on risks to help make underwriters’ jobs easier. The result is a streamlined process that adapts to an insurer’s existing workflow, breaking down any obstacles to exposing unnecessary risk.
Othmane Izi is a product manager for Shift's Underwriting Risk Detection solution.
For more information on Shift Underwriting Risk Detection, request a demo today.