Developing a Framework for Insurance Underwriting Expert System

The process of insurance underwriting determines whether insurance company will accept an application for insurance. It is a complex decision-making process and due to information overload problems, it becomes harder for the insurance companies to efficiently estimate risks in a reasonable time. In this paper we present a framework of an expert system for insurance premium assessment. It combines various Artificial Intelligence techniques, both supervised and unsupervised learning. The proposed framework certainly does not pretend to replace a human underwriter by an electronic one. Rather, it aims at producing viable estimations regarding the clients risk levels, allowing to speed-up the underwriting process and to utilize new potentials.