From tribal knowledge to instant answers: Building Reffy on Databricks
How Databricks built Reffy to turn customer references into an instant, searchable knowledge base using RAG and the Data Intelligence Platform.
Reffy is an agentic application built on Databricks that turns scattered customer reference material into a single, searchable knowledge base. Instead of hunting across YouTube, case studies, and internal decks, employees ask questions in plain language and get answers backed by real customer stories.
The system combines RAG, vector search, and Lakebase to make thousands of customer references discoverable. Since launch, usage has grown quickly: more than 1,800 users and thousands of queries in the first months. The result is more consistent storytelling, faster campaigns, and proof that tribal knowledge can be turned into a governed, scalable asset.