Databricks Genie becomes an autonomous AI coworker for long-running enterprise work

Databricks introduced the new Genie on November 11, 2026: an autonomous AI coworker that takes responsibility for long-running enterprise work instead of answering one question at a time. You reach Genie through Genie One, email, Slack and Microsoft Teams, give it a job and the tools it needs, and it keeps working continuously, tracking progress, following up on missing inputs and acting as new information arrives. Jobs run on a schedule or on triggers, such as a metric dipping, and Genie can run many jobs at once while you supervise, steer or stop them from one control panel. You decide what it may do on its own and what needs your sign-off, within your company's guardrails.

The autonomy runs on context and controls. Genie Ontology gives Genie a governed map of the business, its definitions, metrics, relationships and organizational knowledge, and the new Genie Search searches the lakehouse together with Google Drive, SharePoint, Jira, Confluence and Workday, enforcing each source's own access controls. Genie Agents bring the same setup to whole teams in marketing, sales, finance and operations, and Genie plugins carry the business context into ChatGPT, Claude, Microsoft Copilot Cowork and Grok Bot. Unity Gateway and Unity Catalog govern tool access, trace actions and set spending limits. Databricks says thousands of customers already use Genie in production, naming Zillow, Unilever, adidas, Atlassian and AstraZeneca.

In a benchmark Databricks built with Turing, Genie passed 91% of task runs on business questions in a live Databricks workspace, against 45% and 58% for two anonymized frontier agents, while using fewer than half the tokens; Databricks cautions that the test reflects an established deployment with accumulated business knowledge. Those are the vendor's own figures, not an independent ranking.