Particle.news

Databricks Debuts Agentic CDP and Real-Time Lakehouse Tools

The company says these Azure releases let autonomous agents operate on governed, zero-copy customer data to speed personalization and control AI cost and behavior.

Overview

  • Databricks announced the products at its Data + AI Summit on June 16, unveiling CustomerLake as a lakehouse‑embedded agentic Customer Data Platform now in private preview.
  • CustomerLake uses Profile Agents to build unified Customer 360 records and Campaign Agents to automate audience segmentation, activation, and continuous personalization inside the governed lakehouse.
  • Azure Databricks introduced an LTAP architecture including Lakebase, a managed serverless Postgres with copy‑on‑write branching, and Lakehouse//RT for sub‑second analytics to support live agent workloads.
  • The Genie suite adds embedded AI coworkers and low‑code tools with Microsoft integrations such as Genie for Teams and M365 Copilot (beta), an Excel add‑in (public preview), and a SharePoint connector (beta).
  • Databricks expanded governance with a Unity Catalog AI Gateway, contextual runtime policies, agent tracing, and unified spend controls so enterprises can monitor actions, limit costs, and audit sensitive data access.