Enterprise

Databricks pitches always-on Genie for monitoring business signals

Marketing visibility and customer usage are targets, but the new autonomous mode is still headed for preview and beta.

Photo: Wikimedia Commons / Databricks Inc., public domain — Databricks logo with its stacked geometric symbol and company name

Databricks announced second-generation Genie on November 11, 2026, pitching autonomous business monitoring. Always-on mode isn’t generally available: it’s headed for preview and beta, alongside Genie Search and Ontology Manager.

Databricks proposes monitoring generative engine optimisation (GEO), suggesting content, flagging customer-usage changes and producing weekly executive briefs on revenue trends and feature requests. Running on its own compute, scheduled or triggered, Genie delegates parallel tasks to subagents and surfaces decisions through a priority inbox and email. Each alert starts an email thread; updates stay there.

Genie and Genie Ontology are generally available to Databricks customers. The ontology maintains definitions, metrics, people and processes, learning from dashboards, queries, notebooks and external business-intelligence tools. An organizational connector adds reporting relationships; Ontology Manager flags frequently asked concepts needing clearer definitions.

Genie Search spans more than 100 connectors, including Google Drive, Jira and Confluence. Teams can ingest through Lakeflow Connect, index in place or query applications live through Model Context Protocol (MCP).

“We wanted to boost performance, so we built Genie Search to index those systems as well.” Ken Wong, Databricks’ senior director of product management, told TechTarget.

Slides, dashboards and interactive reports update with governed data. Genie Agents handle PDFs, contracts and slides in Unity Catalog volumes alongside governed tables, without a separate vector index. Actions follow Unity Catalog permissions and Unity Gateway approvals and budgets, with audit trails. Customers pay by usage, not seats; administrators can cap team or user spending.

HelloFresh’s associate director of data science and engineering, Alan Khalil, reported 75% less time identifying winning Meta advertisements and improved programmatic advertising returns. No returns percentage was disclosed; this isn’t evidence of forthcoming always-on results.

“We expanded self-service analytics by putting plain-language access to our data in the hands of teams across the business, not just analysts,” said Marlissa Wong, Headspace’s director of data engineering. Headspace used Genie Agents with agreed business definitions.

TechCrunch’s November 11 report described Databricks CEO and co-founder Ali Ghodsi demonstrating internal Genie: analyzing employee-survey weaknesses, summarizing an unsuccessful sales pitch and required features, and preparing podcast questions from previous appearances.

For small B2B owners, the takeaway is planning, not immediate deployment. A hypothetical monitoring brief could specify an acquisition measure, an attention-worthy change, the responsible person and a decision requiring approval. Search visibility, campaign performance and customer activity are possible starting points. As with scheduled sales-pipeline analysis, the useful question isn’t simply what an agent can watch, but what change warrants action.

Sources