Tyler Cowen of Marginal Revolution has a post talking about how to measure AI progress. He points out that focused applications can often make great progress without the general market really noticing. That is definitely true in the Personal Lines Insurance Risk Selection and Management space. I responded to him with this comment:
I am a founder of Veracityid.com – an insurance risk selection and intervention solutions provider that is adding AI to our Boolean rules engine so that underwriters who understand insurance but not the intricacies of Boolean Logic can define new rules and build the resolution workflow around them. We are also adding AI to our data analytics: identifying anomalies that are associated with lying/fraud and recommending solutions for testing. The great news is that this radically increases the number of ‘experiments’ that underwriters and product people can run with every single experiment directly tied to a reliable and almost instantaneous implementation process. However, the key in a regulated market like personal lines insurance is ensuring that the AI decisioning is tied to very specific actions that can be documented and tested for regulatory compliance. “Black Box” decisioning will never work in regulated environments. Humans must understand that the decisioning process was not only ‘legal’ but also ‘fair’.
AI is rapidly improving the ability of carrier staff to make major bottom line change without needing the costly and time consuming involvement of IT. We’re thrilled to be at the forefront of that process.
