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

Observe makes AI-SRE’s execution identity a monitor-level control

A new GA control shows whose account an investigation uses and lets operators reassign it; the same release also closes a cross-tenant read flaw.

Observe AI-SRE release timeline with identity control, security fix, and rollout dates.
Timeline: dates from the story
By The News Desk· Sep 17, 2026the quick take — two AI hosts go live when you do

Observe has added a small but consequential control to AI-SRE: the monitor form now shows whose identity an investigation runs under, and a monitor editor can claim that identity for themselves. The AI-SRE chatbot can also reassign it. Observe says the change makes clear which connectors and tools the investigation will use and lets users reclaim ownership without engineering help. The feature entered fast-lane general availability on September 16; slow-lane tenants are scheduled to receive it automatically on October 14. Observe changelog, Sept. 16

Why the identity label matters

For an investigation agent, “who is running this?” is not just attribution. Observe explicitly ties the displayed identity to the connectors and tools available during the investigation. That gives operators a visible place to check the authority behind an automated run before treating its output as routine monitor behavior. Reassignment also turns an otherwise implicit dependency on a user account into an editable monitor setting. Observe changelog, Sept. 16

The practical rollout check is straightforward: open each AI-SRE-enabled monitor, record the displayed identity, verify that its connector access is still intended, and assign a durable owner before a user changes role or leaves. Teams should also test reassignment in a non-production monitor and confirm what appears in their audit trail. Observe’s release note does not describe the audit event, credential-refresh behavior or what happens to an in-flight investigation when ownership changes, so those remain questions for tenant-specific validation. Observe changelog, Sept. 16

The same release underlines the boundary

Observe’s September 16 notes also disclose a fixed cross-tenant data-leak bug. A REST read against reference tables or a GraphQL read against monitors could return another customer’s data if a caller guessed or reused a cluster-global numeric ID. The company lists the issue as fixed, but does not state the affected versions, exploitation status or whether customers need to rotate credentials. Observe changelog, Sept. 16

The two changes are separate, but they point to the same operational lesson: agent behavior is bounded by both the identity attached to the workflow and the authorization checks on every data path it can reach. A visible owner is useful only if teams verify that the underlying connectors and APIs enforce the same boundary.

The release also moves AI-SRE in the trace inspector and natural-language semantic search over datasets to slow-lane GA, extending the surfaces where those access checks matter. Observe changelog, Sept. 16

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