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
analysisAnalysis

Oracle Select AI can now call Gemini Enterprise—but the cross-cloud credential becomes an operating dependency

The new provider option reaches Google models through Oracle’s existing NL-to-SQL interface, while OAuth refresh tokens, session policy and regional endpoints create a new failure surface.

Oracle Select AI before and after Gemini Enterprise integration
AI-generated illustration
By The News Desk· Oct 16, 2026the quick take — two AI hosts go live when you do

Oracle has added Gemini Enterprise Agent Platform as an AI provider for Select AI in Autonomous AI Database Serverless. The integration applies to natural-language-to-SQL generation, chat, showsql, explainsql, embeddings and other Select AI capabilities, while preserving Select AI’s existing database-side interfaces. That makes the release more than another model choice: it lets an Oracle-resident text-to-SQL workflow use models available to a customer’s Google Cloud project and region. Oracle release note

The provider changes; the database contract does not

Select AI profiles define the database objects exposed to natural-language queries and can use table names, column names, data types and comments as metadata. Oracle’s profile mechanism also covers SQL generation and execution, SQL explanation, chat, retrieval-augmented generation and embeddings. The Gemini Enterprise option therefore sits behind an existing profile boundary rather than requiring teams to rebuild their database-facing application around a separate Google agent endpoint. Oracle profile documentation

The supported model surface is narrower than the platform name may suggest. Oracle says Select AI supports published Google Gemini models using the generateContent API and published Gemini embedding models using embedContent. Models served through other inference APIs—including Models as a Service, partner models and open models—are not supported through this integration. The selected region must also support the endpoint and model. Oracle profile documentation

OAuth is now part of query reliability

The practical boundary is authentication. Oracle documents Google Cloud OAuth 2.0 rather than a simple static API key for this provider. An operator creates Application Default Credentials with gcloud auth application-default login, then stores the resulting refresh token, client ID and client secret in the database credential used by the Select AI profile. Oracle explicitly warns that these values must not be committed to source control or included in logs or documentation. Oracle profile documentation

That credential also inherits Google organization session policy. Oracle notes that periodic reauthentication controls can invalidate the refresh token when a session expires, producing an invalid_grant response with an invalid_rapt subtype; restoring service requires running the application-default login flow again. Network policy must allow the regional aiplatform.googleapis.com endpoint, or the global endpoint for multi-region applications. Oracle profile documentation

For production teams, the migration checklist is therefore not just model quality and cost. It should include credential ownership, token rotation and reauthentication runbooks, endpoint allowlists, region compatibility, and monitoring that distinguishes an expired Google session from a bad SQL generation. The release keeps the application interface stable, but it makes cross-cloud identity part of the availability path for every generated query.

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