Case Study: Banco Bradesco achieves real-time CRM transformation and doubles conversion rates with Databricks

Banco Bradesco builds real-time CDP with Spark Declarative Pipelines

Banco Bradesco, one of Latin America’s largest banks, needed a faster and more reliable way to collect and analyze customer data from digital and marketing channels. Its legacy systems and fragmented tools made data processing expensive, unreliable, and too slow for real-time customer engagement, so the bank turned to Databricks and Spark Declarative Pipelines to build a modern in-house customer data platform (CDP).

With Databricks, Banco Bradesco implemented a medallion lakehouse architecture that validates and processes data in near real time, automating ingestion, quality checks, and incremental updates across 14 channels. The results were significant: data latency dropped by more than 94%, total processing time fell from 60–80 hours to just 3.5 hours, key jobs now run in minutes instead of hours, and operations require only two people instead of eight to 15, while the bank has processed 110 billion raw observations since October 2023.

Banco Bradesco

Marcos Abrantes Gomes

Senior Manager