Playbooks · Teams
Databricks Genie never sleeps now
Written November 12, 2026
The second generation of Genie, announced November 11, 2026, stops waiting to be asked. It takes a goal, works in the background on its own compute, and comes back through a priority inbox when a person has to decide something. For a data team lead, the decision is the availability split and the governance story — both are spelled out, and they do not line up the way the headline suggests.
What always-on means here
In Databricks' announcement, Genie now operates in autonomous mode, running in the background using its own compute. You hand it a goal and clear boundaries; it takes ownership of the work, delegates pieces to sub-agents that run in parallel, tracks progress, and follows up when an input is missing. The examples are operational, not chatty: keeping a quarterly forecast on track, watching customer usage patterns and telling the account team about meaningful shifts, tracking revenue trends and sending a weekly brief.
Decisions surface in two places. Genie opens to a priority inbox of items that need attention, so nobody digs through old conversations to find the open question. Away from the app, each new alert arrives as its own unread email, updates thread onto the same conversation, and every email links back into Genie.
The context layer underneath
Every action is grounded in Genie Ontology, which captures how the company works — definitions, metrics, the people and processes behind them — and keeps that current as the business changes, learning from dashboards, queries, notebooks, external BI tools and an organization chart connector. Genie Search gives staff and agents one governed search across the lakehouse and the documents, tickets and chats in apps like Google Drive, Jira and Confluence. Databricks offers more than 100 connectors, and content can be ingested through Lakeflow Connect, indexed where it sits, or queried live through MCP. Ontology Manager is the new surface for governing that context, flagging much-asked concepts that still lack a settled definition.
Who approves what
Genie runs inside a company's existing Databricks environment. Every action stays within Unity Catalog permissions and Unity Gateway approval rules and budgets, and is fully auditable. The FAQ puts the admin levers plainly: admins decide which tools Genie can use, set approval rules and spending caps, and trace every action. Pricing, per the same FAQ, is pay-for-what-you-use rather than per seat, with spending limits settable by team or by user through Unity Gateway.
Limits
- Read the availability line before planning around the headline: Genie and Genie Ontology are generally available to all customers. Genie Search, Ontology Manager and the new always-on Genie are coming soon to preview and beta programs. The flagship behaviour in the announcement is the part that is not GA yet, and no preview dates are given.
- This is an enterprise platform feature that lives inside an existing Databricks deployment, with its permissions and budgets. It is not a tool a small team adopts on its own.
- The productivity claims in the announcement, including a customer figure on ad analysis, are Databricks' and its customers' statements, presented as such.
Related playbooks
- Anthropic's model-protection rule is now in force
- The Copilot app now takes GHES accounts
- Copilot code review now has an API
- GitHub Enterprise now includes Copilot at $0 a seat
Sources
Back to: Teams playbooks hub · Also: Claude Meetings, Sheets and Whiteboards land · Run Claude Managed Agents in your own cloud