Databricks Turns Genie Into Autonomous AI Coworker for Enterprise Work

General Motors has deployed more than 1,000 Genie agents across sales, engineering, supply chain and HR operations, Databricks reported November 11 as it relaunched Genie as an autonomous AI coworker for continuous, long-running enterprise work. The upgrade shifts the assistant from answering questions to owning the work around an analysis: monitoring changes, requesting missing input and acting as new information arrives.

The new Genie can run scheduled or event-triggered jobs, juggle several assignments at once and follow up on its own gaps — while users decide which actions require approval. Databricks says thousands of customers already run Genie in production. Its examples include routing insurance claims that fall outside underwriting guidelines and proposing marketing budget changes for sign-off.

For HR and people-operations teams, the deployment question isn't whether the tool is clever. It's which workflows you're willing to hand to an agent that acts under each user's identity, and how much governed knowledge you actually have for it to draw on.

What's generally available now?

Genie Ontology, now GA, automatically builds reusable knowledge snippets from dashboards, SQL queries and notebooks. Unity Catalog Pages and Metric Views preserve explicitly governed definitions. Databricks' OntoRank system weighs source authority, certification, user relevance, popularity and recency to pick context and resolve conflicting definitions — a direct answer to the duplicate-metric problem that plagues most HR analytics stacks.

Genie Search is also GA for Google Drive, Jira, Confluence and Workday, with SharePoint in beta. It indexes content from those systems, enforces their native permissions before returning results, and can combine the index with SQL and MCP connectors that pull live data and take actions in external tools. Organizational directories — GA now, starting with Workday — add reporting relationships and collaboration signals to help rank relevant knowledge.

Not everything is ready. External BI knowledge sources are in beta with initial Power BI support, and Ontology Manager, the curation and review tool, arrives soon in beta.

Can agents read documents and work with partners?

Genie Agents can now analyze PDFs, slide decks and other documents alongside governed tables in Public Preview, returning page-level citations. Native MCP connectors are GA and let agents create tickets, update records and post messages in external tools.

Two beta features target cross-company work. OpenSharing provides a governed agent preloaded with data, instructions and knowledge across Databricks accounts. Genie Clean Rooms lets recipients ask questions without seeing the underlying data or proprietary logic, with conversation and output-size controls.

What controls do administrators get?

The autonomy comes paired with Unity Gateway controls. Traces capture the question, answer, cited objects, response time and actions, giving administrators a queryable record of every task Genie performs. Admins can restrict MCP tools, require approvals, set budgets and spending caps, and review usage. Genie acts under each user's identity and can prepare an access request, but administrators and data owners keep approval authority.

Genie One, the interface for this work, also builds slide decks, dashboards and interactive pages, with scheduled refreshes and shared skills for repeatable workflows. Databricks says Genie One is free for users through January 31, 2027, with more detailed cost controls promised before that date — a window teams can use to pilot before pricing bites.

How did Genie perform on the benchmark?

Databricks released a benchmark built with Turing to back its context-first pitch. In the company's evaluation, Genie passed 91% of task runs versus 58% for the stronger of two anonymized frontier agents, averaging 431,829 tokens and 100.3 seconds per run. Each agent ran every task five times in the same workspace — a fictional UK energy supplier holding tables, notebooks, dashboards, saved queries and documents, including outdated and duplicated assets.

Databricks itself cautions the test reflects an established deployment with accumulated business knowledge, and that gaps may shrink in new deployments or where table schemas alone suffice. That caveat defines the real enterprise requirement here: broader autonomy depends on governed, accessible institutional knowledge, and administrators will decide which jobs — HR included — Genie is allowed to take on.