Machine Learning Driven Agent and Underwriter Segmentation for a Large Commercial and Personal Insurance Policy Provider

THE BACKGROUND

Association for Operations Research and Development (ACORD) standardized forms are used by the insurance industry to increase efficiency, accuracy, and speed of information processing. Every day there are numerous submissions made to this insurance provider for a quotation on a commercial insurance policy. Each policy type has its own ACORD form. Insured or their agent fills and submits ACORD form to the company in a comprehensive PDF file. Each ACCORD form contains specific information about policy holder, their business and type of coverage needed. These PDF files are converted to XML files with the relevant information. The client wanted to mine this information to extract patterns depicting agent and underwriter behavior. The modeling exercise will help in driving their top line revenue growth and identifying intervention mechanisms to reduce the declination rates of the policies received by the company.