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
releaseRelease

Sigma’s PowerPoint export is a governed snapshot—not a live report

The public beta carries report permissions and row-level security into generation, but the resulting deck is still a point-in-time file with its own distribution risk.

Analytics report versus exported slide deck, live source and frozen snapshot.
Side by side: what changed
By The News Desk· Sep 3, 2026the quick take — two AI hosts, this story only

Sigma has put PowerPoint export into public beta, turning a maintained Sigma report into a formatted .pptx file from the product’s export menu. The important deployment detail is not the one-click workflow; it is where governance stops. Sigma says each export queries the current state of the cloud warehouse, applies the report’s permissions and row-level security, and captures the data visible at that moment.

Generation is governed; circulation is not

That makes the generated deck a governed snapshot. It is not a live connection that keeps rechecking Sigma permissions after download. Sigma’s launch post says a changed source requires another export to put updated numbers into a new deck. For teams running weekly pipeline reviews, monthly closes or quarterly business reviews, that is a useful boundary: authorization determines what enters the file, while ordinary document controls determine where the file goes afterward.

The practical acceptance test is therefore two-part. First, export the same report as two users with different allowed rows and verify that the two decks differ as intended. Second, revoke one user’s report access and confirm that the previously downloaded .pptx is handled by the organization’s document-retention and sharing controls. The second check is an operational recommendation, not a Sigma product claim: a detached file cannot inherit a future policy change unless the surrounding document system enforces it.

The beta has report-side limits

Sigma’s export documentation says the exporter requires an account type with the Share documents permission and report ownership or Can edit access. It also caps each exported table at 10,000 rows and says a report must be published before its first export. The same page labels PowerPoint export a public beta whose interface and behavior can change quickly.

Those constraints argue against treating .pptx as a bulk-data escape hatch. They also make publication state part of the rollout checklist: test draft-versus-published behavior, the 10,000-row rendering ceiling, and the exact identity used to generate scheduled or manual decks before replacing an existing reporting pipeline.

Sigma says the beta is available to customers through their representative, with more export formats and finer layout control planned. For now, the cleanest operating model is simple: keep the Sigma report as the maintained source, regenerate the deck when data changes, and treat every downloaded file as a time-stamped distribution artifact—not as a live governed surface.

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.