Understand this story
On November 11, 2026, Databricks released an updated version of its AI Agent Genie, upgrading it into an autonomous AI colleague that can continuously handle long-horizon enterprise tasks within user-defined guardrails, delivering work through Genie One, email, Slack, and Microsoft Teams. Alongside it, Databricks introduced Genie Ontology to provide business definitions and organizational knowledge, Genie Search to retrieve across the lakehouse and business systems, and Unity Gateway to govern tool access, action tracking, and spending caps. On the same day, Databricks announced the general availability of Genie Ontology, an enterprise ontology layer that agents automatically build from teams' daily work and that humans govern, so that every agent shares the same business understanding; this release adds Genie Search enterprise knowledge retrieval, extending the learned context to BI tools and the organizational directory, and introduces Ontology Manager, while Unity Catalog Domains and Pages move to general availability. In a benchmark built jointly by Databricks and Turing, Genie equipped with Genie Ontology ran 91% of enterprise tasks, 1.6 times that of the strongest frontier agent, while using less than half the tokens.
AI-generated from public reports
Latest development · Report
Databricks Launches a New Generation of Genie, an Autonomous AI ColleagueDevelopments
2 confirmed developments
- Source:Databricks Blog
- Source:Databricks Blog
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 12
Yuchen Jin
Databricks Launches a New Generation of Genie, an Autonomous AI Colleague
Databricks has released a new generation of Genie, positioned as an autonomous AI colleague that resides in the background. The company says it scores 91% on enterprise tasks, with accuracy 1.6 times that of the stronger of two frontier agents, while using less than 50% of the tokens. At its core is the Genie Ontology context layer, which can automatically learn how a business operates from the enterprise's data, queries, dashboards and notebooks.
Databricks BlogFirst-partySelected
Databricks Releases New Version of Genie: An AI Colleague That Can Autonomously Handle Long-Horizon Enterprise Tasks
Databricks has released a new version of Genie, upgrading it to an autonomous AI colleague that can continuously handle long-horizon enterprise tasks within user-defined guardrails, delivering work through Genie One, email, Slack, and Microsoft Teams. It is accompanied by Genie Ontology, which provides business definitions and organizational knowledge, and Genie Search, which retrieves across the lakehouse and business systems, while Unity Gateway governs tool access, action tracking, and spending limits. In a benchmark built by Databricks and Turing, Genie driven by Genie Ontology achieved a 91% task accuracy rate, 1.6 times that of stronger frontier agents, while using less than half the tokens.
Nov 11
Unite.AI
Databricks Upgrades Genie to an Autonomous Enterprise AI Collaboration Assistant
On November 11, 2026, Databricks launched a new version of Genie, positioned as an autonomous AI collaboration assistant capable of continuously carrying out long-running enterprise work. It supports scheduled or event-triggered tasks, asks follow-up questions to fill in missing information, and manages multiple tasks simultaneously, with users deciding which operations require approval. Genie Ontology and Genie Search for Google Drive, Jira, Confluence, and Workday are now generally available, and native MCP connectors are also GA, enabling retrieval of context and execution of operations such as creating tickets and updating records. Benchmarks from Databricks' collaboration with Turing show that Genie achieved a 91% pass rate during task runs, compared with 58% for the stronger of two anonymous frontier agents, consuming an average of 431,829 tokens and taking 100.3 seconds per run.
Databricks BlogFirst-partySelected
Databricks Upgrades Genie One to an Enterprise AI Collaboration Platform
Databricks announced that Genie One has become the unified entry point for enterprise AI collaboration, integrating Genie Search, the organizational directory, and artifacts such as slides, dashboards, and interactive pages that can be generated directly. Genie One reads and writes to systems such as Jira and Asana through more than 100 MCP connectors, and supports saving workflows as shareable Skills. Administrators can use Unity Gateway to view call traces and set MCP policies and budgets. Genie One is free for users until January 31, 2027.
Databricks BlogFirst-party
Databricks Genie Agents Support Unstructured Document Analysis and MCP Connectors
Databricks has updated Genie Agents, which can now answer questions across governed tables in Unity Catalog and thousands of documents, providing page-level citations; this capability is in Public Preview. Native MCP connectors have moved to GA, enabling reading from tools such as Slack, Google Drive, and Jira and performing actions such as creating tickets, updating records, and sending messages; OpenSharing and Genie Clean Rooms are in beta, supporting the sharing of governed agents with external partners and customers. General Motors has deployed more than 1,000 Genie agents, and HelloFresh says ad testing is 4 times faster.
Databricks BlogFirst-partySelected
Databricks announces Genie Ontology is now generally available
Databricks announces that Genie Ontology is now generally available. It is an enterprise ontology layer that agents automatically build from teams' daily work and that humans govern, letting every agent share the same set of business understanding. This release adds Genie Search for enterprise knowledge retrieval, extending the learned context to BI tools and the organizational directory, and introduces Ontology Manager, while Unity Catalog Domains and Pages become generally available. In a benchmark built in collaboration with Turing, Genie equipped with Genie Ontology completed 91% of enterprise task runs, 1.6 times that of the strongest frontier agent, with less than half the token usage.
TechCrunch AI
Databricks updates AI agent Genie to complete multi-step tasks without supervision
Databricks on Wednesday released an updated version of its AI agent Genie, which can execute multi-step tasks that take longer to complete and uses an ontology built by Databricks to obtain information customers store on its platform as context. The company says enterprise leaders can assign Genie a variety of tasks and let it complete them without supervision. CEO Ali Ghodsi demonstrated his own actual internal usage, including analyzing the low-scoring parts of an employee satisfaction survey, evaluating improvement proposals put forward by executives, summarizing a sales pitch that did not go well, and preparing questions for a podcast interview. Ghodsi said companies with more than 150 employees can all derive value from it, and mentioned that usage caps can be set through Unity Gateway and cheaper open-source weight models can be swapped in to control costs.
Heat trend for this story
- Current
- 5
- Period peak
- 39 Nov 12 01:00
- 24-hour change
- -75%
Move the pointer or select the chart to inspect hourly heat. Use the left and right arrow keys with a keyboard.