How to turn a plain-English question into SQL
Updated 2026-07-28
Turning a question into SQL is easy to demo and hard to trust. A generic generator writes SQL for a schema it has never seen, so it guesses at table and column names. An AI data analyst does it against your real schema and shows you the plan before it runs.
Here is how to go from a question to correct, reviewable SQL on your own data.
- 1
Connect the schema the query will run against
Correct SQL depends on knowing the real tables, columns and relationships. Connect your source so the analyst generates against your actual schema instead of a generic guess.
- 2
Describe what you want in plain English
Ask for the result you need. The analyst maps your business terms to the right tables and joins using what it knows about your schema.
- 3
Review the plan and the SQL
See a plain-English plan and the exact SQL. Check the joins and filters, adjust if needed, and approve before anything runs.
- 4
Save definitions so it stays correct
When you correct or clarify a field, save the definition. The analyst remembers it, so future questions produce correct SQL without re-explaining your data.
Why schema-aware beats a generic text-to-SQL tool
A generator that doesn't know your schema will invent plausible-looking column names and silently produce wrong results. Generating against your connected schema — with a semantic layer that remembers what your fields mean — is the difference between SQL that looks right and SQL that is right.
Frequently asked questions
It includes text-to-SQL, but it runs against your real schema, shows a reviewable plan, executes the query, charts the result, and remembers your field definitions — an end-to-end analyst, not just a snippet generator.
Common relational databases and cloud warehouses are supported, and cross-source questions are handled with a federation layer.
Related
Get your end-to-end AI business intelligence now.
Ask your data questions in plain English — no SQL. Connecting a database is the technical part: a read-only user and network access, once. If your numbers live in a spreadsheet, none of that applies — a Google Sheet connects with no read-only user and no open port, and CSV and Excel uploads work on the 14-day Pro trial. The one thing no plan waives is your own AI provider key, and a Google Gemini key is free to create.
Every new account starts on a 14-day Pro trial with no card. After it lapses, the free tier stays free.