Home › Resource library › Case studies › 50 reduced manual uw activity and 25 improvement in data entry by digitally extracting unstructured data using ai › Ml nlp from multi 50% reduced manual UW activity & 25% improvement in data entry by digitally extracting unstructured data Share The Client US Based Personal & Commercial Insurance provider Challenge The new business quote process relied on manual processes to develop underwriting risk analysis. Manual activities included: Review of submission attachments including loss runs, location and fleet schedules, application documents, driver lists, etc. Documentation of key profile criteria such as Public Protection Class, classification code, territory code, etc. Entry of submission data into policy admin or rating system. Solution Implementation of Coforge SLICE (Self Learning Intelligent Content Extraction) solution for data ingestion and extraction Implementation of Terrene Labs (TL) Risk Profile solution for third-party underwriting data Integration of SLICE and TL solutions to produce full underwriting risk profiles Full integration of prior company loss reports (provided by prior insurance carriers) into a full underwriting risk profile Value Delivered 50% reduction in manual Under-wiring activities for each new submission 25% improved accuracy from a reduction in data entry errors Allowed UW to generate more business & time for prospective marketing Enabled insurer to discontinue maintenance of proprietary XML schema in favor of industry standard (ACORD XML)'