Case Study: Frende Forsikring streamlines claims reporting and saves hours with Dataiku

Frende Forsikring Using NLP to Automate Claims Reporting

Frende Forsikring wanted to make claims reporting simple for customers, but its single email inbox had to be manually sorted into four claim units, creating repetitive work and slowing handling. The insurer partnered with Dataiku and used its platform to automate the process, combining email text and attachments for model-based routing.

Dataiku helped Frende Forsikring deploy a BERT model and API that automatically classifies incoming claims emails and forwards them to the right unit, while keeping humans in the loop to improve accuracy over time. The solution now routes emails with 98% to 99% precision, saves an estimated 55 to 75 work hours per month, and supports seven Dataiku APIs in production to streamline operations and improve customer response times.

Frende Forsikring