live wire
▸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 says Cortex Analyst apps keep working, but recommends Cortex Agents; semantic views and verified queries carry overSnowflake Documentation▸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 says Cortex Analyst apps keep working, but recommends Cortex Agents; semantic views and verified queries carry overSnowflake Documentation
nl2sql.ai
analysisUnderreported

Fabric’s lineage API gives agents a dependency map—but read permission is not impact authority

The beta item-relations endpoints expose upstream and downstream dependencies as typed edges, giving coding agents a map for impact analysis while leaving change decisions outside the API.

Dependency graph across Fabric items and workspaces with typed edges.
AI-generated diagram
By The News Desk· Oct 16, 2026the quick take — two AI hosts go live when you do

Microsoft Fabric has added a beta item-relations API that returns an item’s dependency graph in machine-readable form. Fabric’s current feature index positions the preview for impact analysis, catalog work, CI/CD checks and lineage-aware AI agents; the REST reference exposes the concrete contract behind that promise.

What the endpoint returns

A caller can request the upstream relations for a Fabric item and receive three collections: related items, relation edges and referenced workspaces. Microsoft’s sample response connects an item to an “Enterprise Sales Lakehouse” with a typed PushData edge and identifies the workspace that contains the dependency.

That is more useful to an agent than a flat inventory. Before editing a warehouse, notebook or semantic asset, an automated workflow can inspect what the target depends on and identify cross-workspace relationships that deserve review. The companion feature listing says the preview also covers downstream relationships, which is the direction needed for blast-radius checks.

The permission boundary matters

The endpoint requires read permission on the target item and either the delegated Item.Read.All or Item.ReadWrite.All scope. It supports users, service principals and managed identities. Those requirements make the graph available to unattended automation, but they do not make a returned dependency safe to modify.

That distinction should shape agent design. A dependency map is evidence for planning: it can tell an agent which assets may be involved, where they live and how Fabric classifies the edge. It is not an approval to rewrite every connected item, nor proof that the graph captures business ownership, freshness requirements or every downstream consumer.

A practical deployment pattern

Teams can use the API as a preflight gate: fetch upstream and downstream relations, record the graph with the proposed change, and escalate when the response crosses workspaces or includes unfamiliar relation types. Actual edits should still run through the permissions and review process for each affected item.

Microsoft labels the endpoint beta, requires callers to send beta=true, and says it is not recommended for production use. The response schema also warns that additional item and relation types may appear over time. For production-minded teams, that means tolerant parsers, explicit unknown-edge handling and versioned audit records are part of the minimum integration—not optional polish.

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.