Query, filter, and aggregate your data using natural language prompts without writing a single line of code.
The AI SQL Assistant uses advanced Large Language Models to translate your English instructions into valid SQL queries. It automatically analyzes the schema of your connected data sources—identifying column names, data types, and relationships—to generate precise data transformations.
The more specific your prompt, the better the result. You can ask for simple filters, complex math, or group-by aggregations.
Good Examples
Too Vague
As seen in the workspace toolbar, you have two options for how the AI applies the query results:
Generates a brand-new table in your "Tables" list. Use this if you want to keep your original data source visible while exploring a new view.
Applies the transformation directly to the table you currently have open. Ideal for incremental filtering or cleanup steps.
Depending on your project needs, you can switch between different AI processing engines in the top-right of the Assistant panel.
Flash Models: Best for simple filters and rapid iteration.
Reasoning Models (Pro): Best for complex multi-step logic and window functions.
The AI Assistant is aware of all tables within your project. You can ask cross-table questions like:
"Join Table A with Table B on the 'SKU' column and show the variance in price."
Pro Tip: Look at the "Available Tables" indicator at the bottom of the prompt box to confirm which datasets the AI can currently access for your query.
Our prompt engineering team can help you write custom instructions for complex business logic.
Contact Analytics Support