live wire
▸Snowflake’s Cortex Agent quickstart routes metrics and transcripts with prompt rules—but documents no authorization boundarySnowflake▸Oracle Select AI generates SQL/PGQ for property graphs—but errors if one AI profile mixes graphs with tables or viewsOracle▸Domo’s data team standardizes MCP builds with shared scaffolding; beta toolkits inherit user permissionsDomo▸Databricks proposes five outcome signals for agentic BI—and says dashboard counts and license usage miss decision qualityDatabricks▸One vibe-coded database app needed 56+ human test hours and three repair rounds despite AI Studio’s verification claimsMancas & Mancas / arXiv▸BigQuery AI.IF’s default cost mode trains a local model for most rows—and drops row-error thresholdsGoogle Cloud▸AWS’s multi-tenant data-agent pattern gives the broker Lambda no data permissions; Cognito roles carry Lake Formation scopeAWS Big Data Blog▸Amazon OpenSearch’s optimized engine adds nested agent traces, but AWS reserves it for log analytics—not vector searchAWS▸Two interpretation-based provenance semantics cannot express SQL bag difference, new theorem showsKolaitis / arXiv▸Databricks Feature Views reuse one definition for training and live scoring, but online lookup still returns latest available stateDatabricks▸Databricks Genie’s Slack and Teams previews default channel answers to public, even when viewers lack source-data accessDatabricks▸Snowflake’s two-tool GraphRAG agent beat a seven-tool graph setup 70% to 60% on 22 ontology questionsSnowflake Engineering▸Semantix opens an Apache-2.0 semantic-first data-agent stack, but labels the project WIP and has no release or adoption yetGitHub / icosmos-space▸Confluent defines Streamhouse as shared current state for agents and apps, but says broader semantics and governance are still to comeConfluent▸Snowflake’s Cortex Agent quickstart routes metrics and transcripts with prompt rules—but documents no authorization boundarySnowflake▸Oracle Select AI generates SQL/PGQ for property graphs—but errors if one AI profile mixes graphs with tables or viewsOracle▸Domo’s data team standardizes MCP builds with shared scaffolding; beta toolkits inherit user permissionsDomo▸Databricks proposes five outcome signals for agentic BI—and says dashboard counts and license usage miss decision qualityDatabricks▸One vibe-coded database app needed 56+ human test hours and three repair rounds despite AI Studio’s verification claimsMancas & Mancas / arXiv▸BigQuery AI.IF’s default cost mode trains a local model for most rows—and drops row-error thresholdsGoogle Cloud▸AWS’s multi-tenant data-agent pattern gives the broker Lambda no data permissions; Cognito roles carry Lake Formation scopeAWS Big Data Blog▸Amazon OpenSearch’s optimized engine adds nested agent traces, but AWS reserves it for log analytics—not vector searchAWS▸Two interpretation-based provenance semantics cannot express SQL bag difference, new theorem showsKolaitis / arXiv▸Databricks Feature Views reuse one definition for training and live scoring, but online lookup still returns latest available stateDatabricks▸Databricks Genie’s Slack and Teams previews default channel answers to public, even when viewers lack source-data accessDatabricks▸Snowflake’s two-tool GraphRAG agent beat a seven-tool graph setup 70% to 60% on 22 ontology questionsSnowflake Engineering▸Semantix opens an Apache-2.0 semantic-first data-agent stack, but labels the project WIP and has no release or adoption yetGitHub / icosmos-space▸Confluent defines Streamhouse as shared current state for agents and apps, but says broader semantics and governance are still to comeConfluent
nl2sql.ai
analysisOracle AI World

Oracle’s lakehouse semantic layer targets the context gap in enterprise NL2SQL—but it is still a roadmap

The new design combines reusable analytic views, agentic metadata pruning and scoped Iceberg credentials; the two headline context services are promised within 12 months.

By The News Desk· Oct 26, 2026the quick take — two AI hosts go live when you do

Oracle has previewed a business semantic layer and AI Data Catalog for Autonomous AI Lakehouse, aiming to give Select AI and other agents consistent business context across Oracle data, Apache Iceberg tables and external catalogs. The practical caveat is timing: Oracle says both headline capabilities will be available within the next 12 months, and its announcement carries a future-product disclaimer rather than a firm ship date. Oracle announcement

The semantic layer moves joins and metrics out of the prompt

Oracle says the Business Semantic Layer will capture dimensions, measures, metrics, hierarchies, calculations and KPIs in reusable Analytic Views. It can import Oracle Database SQL annotations and use AI to suggest views. The company explicitly positions those definitions as grounding for Select AI, NL2SQL, retrieval and analytic insights. Oracle technical blog

That architecture matters because an enterprise question rarely fails only at SQL syntax. The agent must choose the right tables, apply the organization’s definition of a metric and preserve hierarchy logic. Oracle’s proposal puts those decisions in a governed model that can be reused across requests rather than reconstructed in every prompt. The technical post also says support for Apache Ossie will let compatible semantic models be imported, reducing the need to recreate existing definitions. Oracle technical blog

Catalog pruning narrows what the model sees

The proposed AI Data Catalog can mount Oracle and non-Oracle databases, object stores, Iceberg tables and external catalogs including AWS Glue, Databricks Unity Catalog and Snowflake Horizon Catalog. Oracle describes hierarchical “agentic pruning” that narrows discovery from catalogs and schemas to relevant tables and columns, using lineage, relationships and business meaning. The intended result is a smaller, more relevant metadata context for natural-language analytics and agent workflows. Oracle technical blog

The same control plane is intended to apply role-based, object-level policies across distributed sources. For Iceberg, Oracle says it can vend temporary, table-scoped credentials to Spark, Trino and agentic workflows instead of relying on broad, long-lived object-store credentials. Underlying source controls still apply; Oracle says catalog policies work alongside them, not as a replacement. Oracle technical blog

What teams can evaluate now

Not every part of the announcement is equally prospective. Oracle says Data Studio connects to more than 100 sources and lets teams work through natural language, SQL or Python, while expanded Iceberg support covers SQL reads and writes, time travel and external catalogs. Oracle also reports up to 3× faster execution for Data Lake Accelerator enhancements on a 100 TB TPC-DS dataset, but that is a vendor benchmark rather than independent validation. Oracle announcement

For buyers, the immediate test is therefore not “does the promised semantic layer fix NL2SQL?” It is whether today’s catalog connections, source-level authorization and Iceberg execution work against the organization’s real estate—and whether Oracle later delivers the reusable semantics and pruning contract without changing the governance boundary.

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