databricks-genie

First inspect your Unity Catalog tables with get_table_details, then call create_or_update_genie with a display name, table identifiers, a description of table relationships, and sample questions referencing actual column names. A SQL warehouse is auto-detected if not specified.

Use the ask_genie tool with a space_id and your natural language question. It returns the generated SQL, columns, data rows, and a conversation_id you can pass to ask_genie_followup for context-aware follow-up questions.

Use Genie when the space has curated business logic, certified metrics, or when users want conversational exploration. Use execute_sql directly for simple ad-hoc queries, when you already have exact SQL, or when no Genie Space exists for the data.

Poor query generation usually stems from vague column names, missing table comments, or sample questions that don't reference real columns. Add descriptive names, table comments, and instructions via the Databricks Genie UI to improve accuracy.

Space creation fails with a no-warehouse error. Create a SQL warehouse in your Databricks workspace first, or pass an explicit warehouse_id to create_or_update_genie to target a specific warehouse.