The insurance industry is shifting from traditional to near-real-time data-driven models, driven by consumer demand and the need for efficient data processing. To achieve this, software delivery teams must build and maintain data processing pipelines. A usage-based insurance model using MongoDB and Databricks can revolutionize underwriting, enabling personalized and real-time products. The model involves gathering data from connected vehicles, analyzing it with Machine Learning, and creating a personalized premium for customers. A basic data model includes customers, trips, policies, and insured vehicles, which are stored in three MongoDB collections and two Materialized Views. The data pipeline consists of sample data, daily and monthly materialized views, and a scheduled trigger that executes an Aggregation Pipeline to summarize the raw IoT data. Automated decisions with Databricks involve collecting data, posting it to the ML Flow API endpoint, waiting for the model's response, and updating the 'customerPolicy' collection with the calculated premium.