Examples

Every exchange below is a real prompt and a real answer from a live workspace — numbers, table names, and all. Yours will differ; the shape won't.

You do not need a top-tier model for these everyday tasks. We tested these workflows with a smaller, lower-cost model and got useful results. A stronger model may help with an unusually ambiguous or complex request, but it is not a requirement for exploring a schema, asking for numbers, or checking a stream.

Inspect a schema

You: Describe the orders table on my Postgres demo connection — columns, keys, indexes. Lay it out as a table.

AI: public.orders — 10 columns, primary key order_id, foreign key customer_id → customers. Indexes on order_id (PK) and customer_id. No index on region — add one if you group orders by region.

Ask for real numbers

The assistant runs the query — it doesn't guess.

You: What are my top film categories by rental revenue? (sakila sample)

AI:
Category Revenue
Sports $5,314.21
Sci-Fi $4,756.98
Animation $4,656.30
Joined category → film_category → film → inventory → rental → payment, ran in ~55 ms. Want the query to reuse?

Optimize a query

Open Ask AI from a SQL Console to attach the query you are actually working on, then ask what to improve. For example:

SELECT c.customer_id, c.first_name, c.last_name, COUNT(r.rental_id) AS rentals
FROM customer AS c
JOIN rental AS r ON r.customer_id = c.customer_id
WHERE r.rental_date >= '2005-06-01'
GROUP BY c.customer_id, c.first_name, c.last_name
ORDER BY rentals DESC
LIMIT 20;

You: Explain this query and suggest the safest improvement.

AI: The date filter is the expensive part. Check the plan first; if it scans too many rentals, test an index beginning with rental_date.

AI Chat analysing a query opened from the SQL Console

Troubleshoot a stream

You: My MySQL → PostgreSQL CDC stream stopped overnight. List my streams, check status and recent errors, and tell me what to fix.

AI: Found it — the latest run of 10M-cdc-mysql2pg FAILED with Unknown database 'target', source-side. The CDC source points at a database that doesn't exist on the MySQL server. Fix: create it (or correct the name on the connection), then restart.

Summarize what your streams do

You: Summarize what my streams migrate.

AI: (a table of every config) 6 one-shot loads and 1 CDC stream — a full Sakila dataset replicated across MySQL, PostgreSQL, and CSV. The CDC stream to PostgreSQL is incomplete (stopped mid-run, 3 of 16 tables synced).

Validate across sources

In a SQL console with two or more sources attached — say a MySQL database and a folder of CSVs — AI Chat sees every source and its alias, and writes the cross-source query for you.

You: Is my customers.csv export still in sync with the live customers table? Compare row counts and find rows only in one side.

Full multi-source transcripts: plain-English SQL exploration · Postgres + Parquet validation.

Using AI Chat · Generate SQL
Tools reference — every read-only tool behind these answers.