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
newsGoogle Cloud

Google streamlines BigQuery data-agent publishing into Gemini Enterprise

A new preview path removes manual A2A-card and OAuth setup when analytics agents and Gemini Enterprise share a project and region.

Google’s new same-project publishing path versus the older manual cross-project handoff.
Side by side: what changed
By The News Desk· Sep 23, 2026the quick take — two AI hosts go live when you do

Google Cloud has opened a preview route for publishing BigQuery data agents into Gemini Enterprise through Agent Registry, reducing the configuration work needed to put governed analytics assistants in front of business users. The change appeared in Google Cloud’s September 22 release notes.

What changed

When a BigQuery data agent and a Gemini Enterprise application live in the same Google Cloud project, an analyst can now register the agent during BigQuery’s publishing flow. A Gemini Enterprise administrator can then import it from Agent Registry and use Google-managed credentials, rather than copying an Agent-to-Agent card or manually configuring OAuth client credentials, according to Google’s data-agent documentation.

The shortcut has an important boundary: the data agent’s storage region and the Agent Gateway registry binding must match. Google lists US, EU and Global as the available choices, and says the agent region cannot be changed after it is saved. Administrators must also attach the corresponding Agent Gateway to the Gemini Enterprise application before provisioning the agent.

Cross-project deployments still require the older handoff. If BigQuery and Gemini Enterprise are in different projects, the analyst must copy the A2A JSON card from BigQuery and pass it to the Gemini Enterprise administrator. Google-managed authentication remains available in that path, so teams can avoid creating a client ID and secret even though the JSON exchange is still manual.

Why it matters

The release moves Google’s conversational analytics stack closer to a managed promotion workflow: analysts configure the agent around BigQuery tables, views, graphs or user-defined functions; administrators control how it enters the enterprise agent catalog; and users can reach it from Gemini Enterprise through a dedicated chat, an @mention, or automatic routing from a general analytical question. Responses can stream back as text, Markdown, charts or tables, Google says.

That separation also keeps access controls visible. Creators, owners, editors and users have distinct Gemini Data Analytics IAM roles, while the agent acts with the permissions of the person using it. Publishing does not itself grant access to underlying data.

For teams evaluating the preview, the practical checks are project topology, region alignment and role assignment. Same-project deployments get the cleanest path; organizations separating analytics and application projects will still need a documented A2A-card transfer. Either way, the update targets deployment friction rather than model quality: verified queries, metadata and agent instructions remain the mechanisms for improving answer reliability.

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