Databricks Genie Always-On AI Agent Takes Over Complex Enterprise Tasks

Databricks announced the second-generation always-on Genie on November 11, 2026. The autonomous AI agent runs continuously in the background, owns assigned goals, and handles complex enterprise work without constant human prompting.

The system accepts a goal plus defined guardrails. It then delegates subtasks, tracks progress, requests inputs when needed, and surfaces decisions through a priority inbox and threaded alerts.

Core Features of Databricks Genie Always-On Mode

Genie Ontology reached general availability as the central layer that builds and maintains enterprise context. It pulls definitions, metrics, relationships, people, and processes from Unity Catalog, dashboards, queries, notebooks, external BI tools, and directories such as Workday. The ontology updates automatically as new data sources connect.

Genie Search enters preview and indexes the lakehouse along with more than 100 enterprise systems including Google Drive, Jira, and Confluence. Permissions stay enforced through Unity Catalog, and queries can run over live connections or indexed content via Lakeflow Connect or MCP federation.

Governance relies on Unity Catalog for permissions and lineage plus Unity Gateway for tool access, approval rules, budgets, and full audit trails. Every action remains scoped to the original user identity and logged for compliance review.

Background, Availability, and Next Steps

Earlier Genie versions focused on conversational analytics. The product advanced through Genie One, Agents, and Code releases that reached general availability by June 2026. Thousands of organizations already run those capabilities in production, including Zillow, Unilever, adidas, Atlassian, and AstraZeneca.

Core Genie and Genie Ontology are now generally available. Always-on mode, Genie Search, and the Ontology Manager enter preview and beta programs shortly. Pricing follows a usage-based model without seat licenses, and Unity Gateway provides spend controls.

Databricks plans expanded integrations and additional autonomy features. The company also intends to support broader use through Genie MCP and plugins that connect to external systems such as ChatGPT and Claude. Customer feedback will continue to shape refinements to context accuracy and governance controls.