Fraud has always been a problem for property and casualty (P&C) insurance carriers, but as the customer journey has evolved, so too has fraud. With consumers more empowered to shop for insurance without involving a live agent, there is more opportunity for an applicant to submit false information about the risk being presented. Without an agent verifying information, carriers sometimes find themselves in a more compromised position when underwriting.
Another area of concern for carriers is in the service area. As carriers provide more of a self-service experience to their policyholders, the possibility of fraudsters appropriating a legitimate policyholder’s policy to conduct fraudulent actions increases. Further down the value chain lies the most treacherous area for carriers in dealing with fraud―claims. Many claims are now adjudicated without an actual person inspecting the damage, which could serve as an invitation to perpetuate more fraud.
Most of the fraud detection and prevention solutions that are now available to carriers focus on identifying anomalies in a claim or underwriting application, or connecting the dots among fraudsters to reveal an organized network. Essentially, these solutions try to divine the intent of an applicant or claimant and assign a score that assesses the likelihood of fraud being present in any transaction.
But what if carriers could identify a fraudster before any activity (e.g., policy application, claim submission, or service activity) is undertaken?
There are plenty of device verification solutions that can ensure a device being used to execute a transaction is connected to the appropriate individual. However, these solutions don’t verify that the person behind a device is actually that person. This is problematic since it’s possible for someone to steal or clone a device and conduct business as that device’s owner. To avoid this problem, carriers need an identity verification solution. This is where biometric and behavioral biometric fraud detection and prevention solutions can help.
Biometric (i.e., fingerprint, facial, and voice) fraud detection solutions have been used in a range of industries but have not gained much traction among P&C insurance carriers because of inherent challenges. Carriers do not yet store fingerprints and match them to policyholders, not to mention that collecting fingerprints would be cumbersome. The same holds for facial recognition, although this is a more likely solution for carriers to leverage since it is possible for them to tap into driver’s license databases to access pictures and conduct facial matching via a smartphone camera (though to be clear, this is not happening on a widespread basis).
Carriers do have the ability to capture voices when policyholders call and can then use those recordings to match against subsequent calls to verify a policyholder’s identity. This can prevent a fraudster from taking over an unsuspecting policyholder’s account and making changes to a policy, submitting a claim, or altering payment instructions. While there are often safeguards in place, such as a confirmation email or text to the actual policyholder, to prevent fraudulent changes, those safeguards can be compromised. While voice verification can serve as a first line of fraud defense by confirming that the person on a call is indeed the person connected to the policy, it does have limitations, with the chief limitation being that it does nothing to help with non-verbal interactions.
Dealing with non-verbal interactions requires behavioral biometric fraud detection solutions, which focus on how a person interacts with a carrier by typing information into an app, on a website, or through an SMS chat. These solutions also can assess how people are navigating a website or app to detect hesitance or robotic moves (indicating a bot). These solutions measure these behaviors against expected behaviors of either an existing policyholder or of a pool of policyholders.
To illustrate this concept, these solutions can measure how a verified policyholder has behaved during previous interactions and use that as a benchmark to ensure that the person behind any future interactions is actually the policyholder. For example, these solutions can measure how long a person pauses while entering pertinent information (such as a phone number or an address) and compare it against a benchmark of expected behavior. Any legitimate policyholder would be able to enter this information quickly and without much thought. A fraudster, on the other hand, might have to pause to confirm that the information is correct. If an anomaly is detected, these solutions can alert carriers, which can then decide what action to take (e.g., ask multifactor authentication questions, refer the transaction to a live person, or stop the transaction altogether).
Certainly, there is no shortage of vendors offering biometric and biometric behavioral solutions, and carriers should consider how those solutions can be integrated into their respective fraud detection and prevention efforts. However, carriers must realize that these solutions are limited in that most of them are not able to divine intent or gauge the likelihood of fraud being present in a transaction. Indeed, it is entirely possible that a legitimate policyholder, who would not be flagged by any of these solutions, could perpetrate fraud or be a part of a network that could perpetrate fraud. Still, that limitation should not dissuade consideration.
Add new comment