Stagwell has shared a privacy-safe identity matching pattern on Databricks that helps brands connect fragmented first-party data with an identity graph without sending raw customer records outside their own environment.

What’s new

• Install the app where the data already lives: Databricks Apps let partners deploy secure data and AI applications directly inside a customer’s Databricks workspace instead of requiring a separate environment or data export.
• Keep data governed inside the customer workspace: Databricks Apps can use the logged-in user’s identity for data access, so Unity Catalog permissions, row filters, and column masks still apply automatically.
• Use clean rooms for privacy-safe matching: Databricks Clean Rooms provide a secure environment where multiple parties can work together on sensitive data without direct access to each other’s raw records.
• Protect the app while simplifying deployment: Marketplace distribution gives data providers a way to package and distribute applications and data products more broadly through an open marketplace for data, analytics, and AI assets.
• Share outcomes without copying the source data: Databricks sharing capabilities are built around live sharing with centralized governance, which helps partners deliver results while avoiding unnecessary replication.

👉 Read the full post here