Case Study: Progressive achieves faster telematics modeling with H2O.ai

Progressive turns 14 billion miles of telematics data into faster models with H2O.ai

Progressive, a major auto insurer, faced the challenge of analyzing its immense Snapshot telematics data—over 14 billion miles of driving data—to build sophisticated models for usage-based insurance, churn, fraud, and claims. Their existing basic models could not keep pace with the data volume or the range of questions the analytics team needed to answer. To address this, they turned to vendor H2O.ai and its H2O-3 platform.

The solution implemented by H2O.ai involved pairing H2O with Progressive's existing analytics stack, which included SAS, DB2, SQL Server, and Hadoop. This integration enabled Progressive's team to move beyond basic models and use more sophisticated machine learning to build many models in far less time. The measurable results were significant: the Snapshot program analyzed over 14 billion miles of data, returned more than $500 million in discounts to customers, and provided average savings of 10-20% for participants.

Progressive