Claims Forecasting & Databricks Lakehouse for an Insurer
Insurance: Claims data sat in legacy warehouses; actuarial and finance teams waited weeks for reserving and claims-volume forecasts. Data Refresh Weekly →…
Client: General Insurance Company. Industry: Insurance.
Challenge: Claims data sat in legacy warehouses; actuarial and finance teams waited weeks for reserving and claims-volume forecasts.
Solution: Migrated claims, policy and payments data to a Databricks lakehouse with Unity Catalog, then built claims-frequency and severity forecasting with MLflow-managed models refreshed monthly.
Technologies: Databricks, Delta Lake, Unity Catalog, MLflow, Spark.
- Data Refresh: Weekly → daily
- Forecast Turnaround: -75%
- Records Migrated: 1B+
- Platform Cost: -30%