live wire
▸Snowflake agent evals can pin judge versions; unpinned metrics may roll forward and shift scoresSnowflake▸ThoughtSpot makes Spotter 3 default for new instances; it requires download rights and always shares query data with an LLMThoughtSpot documentation▸MotherDuck adds image classification and object location to prompt_jev(), returning typed SQL outputs with confidence scoresMotherDuck▸BigQuery conversational analytics adds Montréal and Toronto; mixed-region agents still default to globalGoogle Cloud Documentation▸Fabric enables Azure OpenAI Copilot access by default, but OpenAI-subprocessor and cross-region processing/storage switches remain offMicrosoft Learn▸Snowflake Cortex Agents now continue by default when named tools are inaccessible, emitting warnings instead of failing the runSnowflake documentation▸Databricks Genie One can use group memberships to tailor answers, while Unity Catalog remains the access-control boundaryDatabricks documentation▸Fabric data agents can query Power BI models with Read alone; Build and workspace access are unnecessaryMicrosoft Learn▸Fabric data-agent Git configs expose column flags, but selecting a table still includes every columnMicrosoft Learn▸Google Data Agent Kit offers remote Spark kernels, but notebook exports can still move results into local Pandas or ArrowGoogle Cloud Documentation▸Cube says production analytics agents need explicit partial-result states, caller-filtered model search and durable artifactsCube▸Thinking Machines names Bridgewater behind ReViSQL-K2.6, citing 92.97% SC-16 accuracy at 15% of frontier-model costThinking Machines Lab▸AWS says a mismatched data/tool/prompt release cut its Field Sales agent pass rate from 78% to 38% in pre-productionAWS Artificial Intelligence Blog▸NL2SQL Multi-Agent ships evaluation artifacts across BIRD, CSpider, Archer and KDD DataSpace—but no software licenseGitHub▸Snowflake agent evals can pin judge versions; unpinned metrics may roll forward and shift scoresSnowflake▸ThoughtSpot makes Spotter 3 default for new instances; it requires download rights and always shares query data with an LLMThoughtSpot documentation▸MotherDuck adds image classification and object location to prompt_jev(), returning typed SQL outputs with confidence scoresMotherDuck▸BigQuery conversational analytics adds Montréal and Toronto; mixed-region agents still default to globalGoogle Cloud Documentation▸Fabric enables Azure OpenAI Copilot access by default, but OpenAI-subprocessor and cross-region processing/storage switches remain offMicrosoft Learn▸Snowflake Cortex Agents now continue by default when named tools are inaccessible, emitting warnings instead of failing the runSnowflake documentation▸Databricks Genie One can use group memberships to tailor answers, while Unity Catalog remains the access-control boundaryDatabricks documentation▸Fabric data agents can query Power BI models with Read alone; Build and workspace access are unnecessaryMicrosoft Learn▸Fabric data-agent Git configs expose column flags, but selecting a table still includes every columnMicrosoft Learn▸Google Data Agent Kit offers remote Spark kernels, but notebook exports can still move results into local Pandas or ArrowGoogle Cloud Documentation▸Cube says production analytics agents need explicit partial-result states, caller-filtered model search and durable artifactsCube▸Thinking Machines names Bridgewater behind ReViSQL-K2.6, citing 92.97% SC-16 accuracy at 15% of frontier-model costThinking Machines Lab▸AWS says a mismatched data/tool/prompt release cut its Field Sales agent pass rate from 78% to 38% in pre-productionAWS Artificial Intelligence Blog▸NL2SQL Multi-Agent ships evaluation artifacts across BIRD, CSpider, Archer and KDD DataSpace—but no software licenseGitHub
nl2sql.ai
newsENTERPRISE AI

Databricks puts Genie inside a new account-level Governance Hub

The beta gives platform teams one place to inspect data coverage, AI usage and cost — while keeping policy-changing actions on the roadmap.

Before-and-after Databricks governance view with scattered records versus one unified hub.
Side by side: what changed
By The News Desk· Aug 26, 2026the quick take — two AI hosts go live when you do

Databricks has launched Governance Hub in beta, adding an account-level control surface for data governance, AI activity and platform cost across AWS, Azure and Google Cloud. The company says the new hub is designed for teams overseeing hundreds of workspaces, where the evidence needed for an audit or cost investigation is otherwise spread across system tables, workspace views and third-party tools. Databricks announced the beta on August 26.

What administrators get now

Governance Hub is split into Data, AI and Cost views. The Data view highlights missing tags, ownership, descriptions and classification coverage; its Access Insights view lets administrators inspect direct and inherited access for a user, group or service principal. The AI view reports token consumption, model activity, per-user spending and guardrail coverage from Unity AI Gateway, while Cost surfaces spend trends and resources missing allocation tags. The product documentation describes these as account-level views with drill-downs.

The practical change is consolidation rather than a new permission model. Databricks says the hub respects existing Unity Catalog permissions and introduces no new access controls. Account administrators receive the broadest view; workspace administrators see cost information for their assigned workspaces, and metastore administrators see data information for their metastores. The documentation also warns that data can take up to a day to appear after an administrator enables the preview. Those role boundaries and the loading delay are documented here.

Genie moves from analysis to governance questions

The release’s text-to-SQL angle is the integration with Genie. Administrators can ask natural-language questions such as why costs spiked or which sensitive tables lack masking policies, with answers grounded in the hub’s governance data rather than a separately maintained dashboard. Databricks says Genie understands the context of the Data, AI and Cost views, so users can investigate token usage or unclassified assets without writing queries. The launch post gives those examples explicitly.

For practitioners, that makes metadata quality part of the query interface itself: if ownership, tags or classifications are incomplete, the hub is intended to expose the gaps before teams rely on an agent’s answer. It also gives operators a common place to watch who is using models and what that traffic costs, rather than treating conversational analytics as an isolated front-end feature.

There is an important boundary in the beta. Genie can answer questions and recommend next steps today, but Databricks describes policy configuration, alert creation and implementation of recommendations as capabilities coming later. Governance teams should therefore evaluate the preview as an observability and investigation layer, not yet as an autonomous remediation system. Databricks places those actions on its roadmap.

sources

Filed by The News Desk. Corrections: desk@nl2sql.ai · Our standards →

comments · 0

    Comments are moderated before they appear. Your email is used once to confirm it is you — never shown, never sold. Corrections and questions get an answer from the desk when we have one.

    more from the desk