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
guideDocumentation gap

Databricks lists Genie chat downloads—but its linked guide does not yet explain them

The Oct. 1 release note names a portable chat export; the linked operating guide still documents only pinning and renaming conversations.

Release note promises chat download; guide still only shows pin and rename.
Side by side: what changed
By The News Desk· Oct 5, 2026the quick take — two AI hosts go live when you do

The release note and the missing procedure

Azure Databricks’ Oct. 1 AI/BI release notes say Genie One now streams responses in real time and lets users download a chat. The same entry directs readers to the product’s chat-management guide for the download feature. Databricks also cautions that releases are staged and an account might not receive an update until a week or more after the initial release date. Azure Databricks release notes

That linked guide, also dated Oct. 1, does not currently describe a download control, file format, included fields or permission behavior. Its Manage chats in the sidebar section documents two actions: pinning a chat and renaming it. A search of the fetched guide exposed no download instruction. Azure Databricks chat guide

This does not show that the feature is unavailable. It shows that the release announcement currently carries more operational detail than the linked procedure.

Why the gap matters

A downloadable analytics-agent conversation can become a useful review artifact. A team could attach it to an incident, hand it to an analyst, or retain the question-and-answer sequence behind a decision. But those uses depend on details the linked guide does not yet specify: whether the export includes generated SQL, citations, visualizations, timestamps, user identity, hidden instructions or only visible chat text.

The guide does document adjacent boundaries. Genie One first looks for a relevant Genie Agent and otherwise searches dashboards, queries and metric views. Users need CAN USE on at least one SQL warehouse. The page also says citations identify knowledge sources and that scheduled tasks email results with visualizations and a PDF attachment. None of those statements defines the new manual chat download. Azure Databricks chat guide

What administrators should verify

Before treating a downloaded chat as an audit record, test the feature in the deployed workspace and answer four questions:

  1. Scope: Does the file contain the full thread or only the selected answer?
  2. Evidence: Are citations, generated queries and visualizations preserved?
  3. Identity: Does the export record who asked, generated and downloaded the content?
  4. Permissions: Can a downloaded artifact expose data after the recipient loses access to the original source?

Streaming responses create a second review distinction: text shown while generation is in progress should not be treated as the final answer merely because it is visible. The release notes promise progressive display, but do not describe whether intermediate text can change or how an export handles it. Azure Databricks release notes

The practical rule is simple: treat the release note as evidence that Databricks lists the capability for staged rollout, not as a complete retention or audit contract. Until the operating guide catches up, validate the exported artifact directly before building review, compliance or incident workflows around it.

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