Nov 4, 2025 – Ask AI Accuracy, AI Agents for Governance, and more!

October's update is all about making your data work smarter, not harder. We've supercharged Ask AI with better analysis, live data querying, and Claude Sonnet 4.5. Plus, we're expanding AI Agents beyond tagging to help automate governance workflows. With improved semantic models and OpenLineage support, this release makes it easier to trust your data and scale your catalog 🤓.

✨ What's New

🧠 Ask AI - Smarter Analysis and Live Data Querying

We've made major improvements to Ask AI to handle more complex workflows and deliver better answers:

Unified Search & Ask AI Search and Ask AI are now one experience. Start from the homepage and get full AI-powered answers, not just search results. No more switching between tools.

Smarter Responses

• Claude Sonnet 4.5: Upgraded to the latest model for more accurate answers

• Follow-up questions: Continue conversations naturally without starting over

• Quality improvements: Better handling of complex queries across the board

AI Admin Rating (Private Beta): Admins can review and rate AI responses to improve quality over time. Learn more in our docs.

Live Data Querying (Private Beta) Ask AI can now query your actual data sources directly, not just metadata. Ask questions that require real data analysis and get answers backed by live query results.

👉 Interested in AI Admin Rating or Live Data Querying? Reach out to our Customer Success team.

🔗 OpenLineage Integration (Private Beta)

We now support OpenLineage for ingesting lineage from data pipelines and processing frameworks. This gives you end-to-end visibility across your entire data stack, from raw sources to final dashboards.

• Ingest lineage from Airflow, Spark, dbt, and other OpenLineage-compatible tools

• See how pipeline runs connect to warehouse tables and BI assets

👉 OpenLineage is currently in private beta. Learn more in our OpenLineage docs.

✨ Auto Description Generation

Select Star can suggest descriptions automatically whenever your data assets lack documentation.

• No more blank dashboards, tables, or columns.

• Just open any asset in Select Star and AI-generated descriptions will appear automatically.

• You can now enable this feature directly in AI Settings

➕ Improvements

🤖 AI Agent - Automate Your Governance Workflows (Private Beta)

Building on our AI Tagging Agent, we're expanding AI Agents to help automate more of your governance work:

• Automated tagging: Use natural-language rules to tag assets automatically

• Owner management: Let AI suggest and assign owners based on usage patterns and metadata

• Collection management: Automatically organize assets into collections based on your rules

AI Agents run continuously in the background, keeping your catalog clean and consistent with less manual work.

👉 Check out AI Agent docs and reach out to our Customer Success team for early access.

🧠 Semantic Models - More Complete and Connected

We've added key improvements to Semantic Models:

• Joins Support Define and visualize joins between entities directly in your semantic models. This makes relationships clearer and models easier to query.

• Looker Metrics Generation Generate metrics from Looker automatically, pulling in dimensions, measures, and relationships to build a complete semantic layer.

• Example Queries

Each semantic model now includes example queries sourced from actual warehouse usage. Select Star automatically identifies and ingests popular queries, filtering out test queries and noise. This helps you understand how tables are actually used in practice and get started faster with real-world examples.

Learn more in our Semantic Models docs.

• Semantic Models MCP Tools

We've added semantic models support to the MCP server. AI agents and tools can now access your semantic models to answer questions more accurately and build better queries.