Case Study: Optum achieves 2x faster claims processing and millions in recovered revenue with Databricks

Optimizing healthcare insurance with data and AI

Optum, a healthcare services company, helps hospitals detect and recover underpaid insurance claims but was constrained by a legacy Oracle-based analytics system. With millions of claims submitted daily and an 8-hour SLA to return recoverable revenue, the rigid, costly infrastructure couldn’t scale compute independently from storage or process data fast enough, risking millions in lost hospital revenue.

Moving to Azure Databricks with Delta Lake and Apache Spark simplified infrastructure, automated cluster management, and enabled blended batch and streaming pipelines. Pipeline performance roughly doubled—processing 80 million claims in 86 minutes (≈1M claims/min) versus 180 minutes previously—allowing faster model training and delivery of results to hospitals and preventing millions of dollars in potential revenue loss.

Optum

Mohammed Salim Sayed

Principal Data Engineer