AI Agent SQL for Google Sheets/CSV in PostgreSQL (n8n)

Ask financial questions in plain English—then get exact results from Google Sheets/CSV using PostgreSQL + an AI SQL agent

This n8n workflow lets you query structured financial data from Google Sheets or CSV files by converting natural-language questions into SQL with an AI agent. Instead of relying on vector databases (which often struggle with numerical accuracy), it uses PostgreSQL for efficient storage and fast, precise numeric querying.

What this workflow does

• Retrieves your data from a Google Sheet (via a public/shared URL) or from CSV.
• Infers the data schema and creates a corresponding PostgreSQL table.
• Populates PostgreSQL with your sheet/CSV records.
• Uses an AI agent to translate your natural-language question into optimized SQL.
• Executes the SQL in PostgreSQL and returns precise numerical results quickly.

Use cases

• Finance and ops reporting: “What were total expenses by category last month?”
• Sheet-to-database querying: Turn a structured Google Sheet into a queryable PostgreSQL dataset.
• Non-SQL teams: Let SaaS operators and automation engineers generate correct SQL without writing it manually.
• Cost-effective analytics: Avoid vector-database inefficiencies while keeping queries performant as data grows.

Technical details

• AI-generated SQL driven by an agent to create queries from natural language.
• Database: PostgreSQL (standard installation; no special extensions required).
• Integrations / nodes: google sheets, postgres, code, set, if, and sticky note.
• Template walkthrough: Watch the full breakdown here: 📺 Video Link.

Why it’s useful: No SQL knowledge required, more reliable numeric results than vector approaches, and designed to scale without overwhelming the language model. Fully free and open-source.