Quick Answer

Describe your query intent in plain English alongside your schema, and AI will write the SQL. For optimization, paste EXPLAIN ANALYZE output and ask for index and rewrite suggestions.

Always provide the schema — without it AI guesses column names
Verify on a staging DB before running in production
AI is excellent at rewriting CTEs and window functions

What You'll Need

A relational database (PostgreSQL, MySQL, SQLite, SQL Server)
Your schema as DDL or an ERD
An AI chat or IDE tool
A SQL client (psql, DBeaver, TablePlus)

Steps

Paste the schema first. Export with pg_dump -s or copy CREATE TABLE statements.
Describe the query in plain English. Example: Find top 10 customers by revenue in Q3 2026 who haven't purchased in the last 30 days.
Review the generated SQL. Check JOINs, WHERE filters, and aggregation grouping.
Run on a non-prod copy. Never run AI SQL directly on production.
For optimization, provide EXPLAIN. Run EXPLAIN (ANALYZE, BUFFERS) SELECT ... and paste output. Prompt: Suggest indexes and query rewrites.
Apply suggested indexes carefully. CREATE INDEX CONCURRENTLY in Postgres to avoid table locks.
Compare runtimes. Before/after EXPLAIN ANALYZE — expect 10-100x speedups for missed indexes.

Common Mistakes

Not providing the schema. Hallucinated column names lead to runtime errors.
Accepting SELECT *. AI loves wildcards. Ask for explicit columns.
Ignoring N+1 potential. Review whether the query runs in a loop.
Skipping LIMIT in exploratory queries. Expensive scans on big tables.

Top Tools

Tool	Strength
ChatGPT / Claude	General SQL generation
GitHub Copilot	Inline in `.sql` files
SQLAI.ai	Dedicated SQL assistant
pganalyze	Postgres query advisor
EverSQL	Index advisor for MySQL

Conclusion

AI turns SQL from a gatekeeper skill into an accessible tool. Pair it with EXPLAIN ANALYZE and never skip the staging run. Misar Dev has a built-in SQL workspace with AI query writer.