Databricks

Databricks is the lakehouse layer underneath a growing share of modern data stacks — Delta tables governed by Unity Catalog, queryable through SQL warehouses, served to BI and notebooks alike. When the integration lands, any Unity Catalog table reachable from a SQL warehouse becomes a live source for canvases, functions, and chats.

What you'll be able to do

• Pull governed lakehouse data into a chat — Unity Catalog's row- and column-level grants carry over, so the user only sees what they're already entitled to.
• Run SQL against Delta tables for the kind of large-scale analytics that doesn't fit in a transactional database.
• Cross-source joins in a canvas — Databricks tables alongside Snowflake, BigQuery, or Postgres in a single view.
• Schedule functions against the lakehouse — daily aggregates, weekly summaries, anomaly detection runs that post results back to a chat.
• Schema introspection at scale for the catalogs and schemas your team actually has access to.

How it'll work

Connect via Databricks Personal Access Token or OAuth to a SQL warehouse. myai talks to the warehouse using the standard Databricks SQL connector — same protocol the official drivers and dbt use. Unity Catalog grants are the source of truth for what's queryable.

Tell us you want Databricks

Use cases we're hearing

• Lakehouse-backed reporting delivered to non-SQL users through chat instead of through a separate BI seat.
• Cross-warehouse joins for orgs running a mixed Databricks + Snowflake or Databricks + Fabric stack.
• ML feature inspection — pull feature-store rows directly into a conversation for ad-hoc debugging.

Reference

• Connection model: Databricks SQL warehouse via PAT or OAuth
• Unity Catalog grants enforced — no separate permission layer
• Scope: ~40 operations spanning SQL execution, Unity Catalog introspection, and warehouse/cluster status