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
guideMigration guide

Microsoft is retiring Foundry’s direct Fabric data-agent tool—and the replacement changes agent routing

Teams have until November 6 to move to Fabric IQ, gaining multi-agent catalog discovery and MCP endpoints while keeping cross-geo and identity constraints.

Timeline of the Foundry-to-Fabric IQ migration deadline and rollout steps.
Timeline: dates from the story
By The News Desk· Oct 16, 2026the quick take — two AI hosts go live when you do

Microsoft has put a deadline on one of the integration paths between Fabric data agents and Microsoft Foundry. The company’s documentation now says the direct Microsoft Fabric tool will stop being supported on November 6, 2026, and tells customers to migrate affected Foundry agents to the Fabric IQ (OneLake Catalog) tool before then.

This is not merely a renamed connector. The retiring path attaches one Fabric data agent to a Foundry agent by supplying a Fabric workspace ID and artifact ID. It uses end-user identity passthrough, but Microsoft documents a one-data-agent limit for each Foundry agent.

What changes with Fabric IQ

The replacement exposes published Fabric data agents through the OneLake Catalog. In the Foundry portal, builders select agents by name from the set they can access instead of manually looking up IDs. One Foundry agent can call multiple Fabric data agents, and each appears as a Model Context Protocol tool.

That changes the operating problem. A team can split sales, supply-chain and support data into separate agents and let a Foundry agent choose among them. Microsoft explicitly warns, however, that instructions must say when each data agent should be used; without clear routing guidance, the Foundry agent might skip a data agent or call the wrong one. Migration testing therefore needs to cover not only whether each agent answers correctly, but whether the orchestrator selects the intended agent—and whether questions spanning domains invoke the right combination.

The catalog also adds governance signals to discovery. Its list shows each published agent’s workspace location, endorsement and sensitivity label. Draft agents do not appear, and users see only agents shared with them. Requests run under delegated user authentication, so Fabric applies the signed-in user’s data permissions.

The controls teams still need to check

Fabric IQ does not remove the cross-boundary caveat. Microsoft says responses from a Fabric data agent can be sent outside Fabric’s compliance boundary or geographic region and then processed or stored under Foundry’s terms. Cross-geo processing and storage settings must be enabled where required.

The replacement also has deployment prerequisites: Fabric IQ is unavailable in regions where Power BI is the only Fabric workload, and the Fabric and Foundry agents must be in the same tenant. The portal workflow requires a Foundry Project Manager to create the Fabric IQ connection, while a Foundry User can build and test an agent in an existing project.

For migration owners, the practical checklist is short but consequential: inventory Foundry agents using the retiring Fabric tool; replace each connection with Fabric IQ; republish and share the relevant Fabric data agents; rewrite routing instructions for multiple tools; verify delegated permissions and regional settings; then test each data agent individually and with cross-domain questions before November 6.

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