The new Genie: autonomous AI for complex enterprise work
Databricks introduces Genie, an autonomous AI coworker designed to handle long-running enterprise tasks continuously, within set guardrails, using business context and controlled access tools.
Genie operates as an autonomous AI coworker that manages ongoing enterprise tasks such as monitoring pipeline performance or adjusting budgets, tracking progress and following up on missing inputs without requiring constant human oversight. Employees set guardrails, supervise progress, and approve key decisions while Genie handles the routine work, adapting to feedback and checking in when input or approval is needed.
Genie relies on Genie Ontology, an agent-built and human-governed map of business definitions, metrics, and relationships, to provide trusted context for its actions, while Genie Search retrieves information across the lakehouse and connected business systems to ensure accurate and complete responses.
Unity Gateway allows administrators to control tool access, trace actions, and set spending limits, ensuring Genie’s autonomy remains within organizational policies, while Genie can work across multiple jobs simultaneously with supervision and control managed through a single interface.
Genie expands capabilities with new features such as unified search across documents and systems, creation of shareable artifacts like slide decks directly in chat, and cross-application collaboration via MCP connectors to tools like Slack, Microsoft 365, and Jira, all grounded in the same business ontology.