As research for the eBook, 3+ Keys to Proactive Underwriting, Cogitate’s team spent time with John Petricelli, Chief Data Officer of our ecosystem partner Confianza. Integrated with Cogitate’s DigitalEdge Platform, Confianza helps insurers accelerate digital transformation. One of their trending use cases is the identification and prevention of premium leakage and fraud. Confianza’s risk intelligence identifies key policyholder motivations and behaviors as well as scoring/flagging of submissions by risk appetite and selection criteria.
Combining over 50 accurate consent-compliant and non-FCRA data sources with submission data such as driver name, address, and birth date, Confianza’s proprietary risk modeling detects undisclosed drivers, false garaging, likelihood to exaggerate claims, prior vehicle damages, synthetic identities, and more.
At Confianza, John Petricelli has engineered the intelligence that produces a 360 view of your prospects and policyholders with underwriting quality, to make smart decisions for your company and offer your policyholders the best possible coverage. Our conversation is summarized to provide you with an overview of the how, what, and why of Confianza’s risk intelligence and the benefits brought to underwriting auto lines.
How vast is the database?
Confianza has been nicknamed by its clients as “the database of America.” The data includes every adult, every household, and every property in the US. With just a name, address, and date of birth, Confianza identifies all vehicles (VIN), members of the household, and much more. Confianza also maintains data on every business in the US and every commercial property, by type.
How does the risk modeling work?
The power of the data is in the proprietary engineering intelligence that interconnects and gives rise to real insights about people and their assets. For example, a very detailed utility that looks at addresses with 29 different conditions will trigger address exceptions, while some competing providers will not flag the error or omission which could lead to hidden, material risk indicators. Confianza will flag that submission as a failed lookup, signaling a requirement for corrective action.
The data intelligence will confirm the proper identity of the consumer presented, the territory, address, risk exposure, the asset – its current condition, use of the asset, details of the driver(s) presented – and uncover those not disclosed, rate variable integrity for all household drivers and share a motivation field, which indicates financial capability and the likelihood of purchase. The auto details also include full registry ownership, title status, and events. Confianza links this information to the people and households. It also confirms that the current personal or business use of the vehicle matches the registered use.
All of these factors provide the transparency required to evaluate price risk based on a source of truth that fact-checks possible omissions, unintended errors, and concealment of information that can lead to losses.
The fight against premium leakage
The first step Confianza takes with many clients is to scan their book for premium leakage based on a very conservative model. Most often, discovery finds undisclosed drivers account for a minimum of 15% premium leakage. Non-standard auto can be upwards of 25%. On-going book monitoring is extremely important as Confianza has found at least 15% of policies have a material change before renewal, which may not be captured at renewal if reliant on the policyholder to disclose the change. This illustrates the importance of data validation at renewal as well as submission.
Streamlining the underwriting process
Integration of Confianza with Cogitate DigitalEdge Policy streamlines the underwriting process by providing a digital 360 view by flagging the high conversion leads, eliminating the early no-goes, and reducing the time discussing exception handling with agents from an average of 20 minutes to 2-3 minutes. Insurers can screen submissions against very specific underwriting criteria for early decisions, saving time and investment in additional data. For example, if the model identifies a submission as a synthetic identity, the process stops before any additional data is called.
For more information on our technology ecosystem, visit our partner page here.