An Expert System Using Nonmonotonic Techniques for Beneefits Inquiry in the Insurance Industry

This paper describes BenInq, an expert system for beneets inquiry in the insurance industry. BenInq is designed to be used by both customer service representatives, who answer customers' questions in real time, and policy modiiers, who update insurance products. The main challenges were the design of the KR structure | in particular, representing a huge knowledge base at varying levels of granularity, and the development of the reasoning methods | in particular , developing a method that determines which insurance regulations apply to an insurance service. This required nonmonotonic reasoning , which we modeled using an extension of inheritance methods. BenInq represents one of the few large scale industrial applications that explicitly uses formal nonmonotonic reasoning techniques.