AI Database Chat Agent Pro

Who is this for?

You run ops for a 5-person store fielding 8 order, stock and refund chats a day, and this Postgres chat workflow answers from live tables. Each request needs the right join, a Sheets cross-check, and a grounded reply in chat.

What problem does this solve?

You open the Postgres client, run the join, then copy the row into Slack. When a chat message arrives, the agent queries Postgres and Sheets, logs the query, and drafts the answer. You just check the draft instead of writing SQL from scratch.

What this workflow does

1. Postgres (Postgres)
2. AI Agent (AI Agent)
3. OpenAI Chat Model (OpenAI)
4. Google Sheets - Query Log (Google Sheets)
- Includes Daniel Grey pack: Slack - Error Alerts, Wait - Rate Limit, Set - Audit Trail (all wired).

How it works

A chat message kicks it off → AI Agent, Set - Audit Trail, Google Sheets - Query Log, ending in Wait - Rate Limit.

Setup

1. Open danielgrey.shop/app/templates/dg-0001 → Copy JSON → n8n → Import from Clipboard (or Push to n8n)
2. Connect credentials: Google Sheets, OpenAI, Postgres, Slack
3. Run once manually to verify each branch — it imports clean and goes green first try — then a 10-minute review and activate.

Requirements

- n8n 1.50.0 or later
- Credentials for Google Sheets, OpenAI, Postgres, Slack
- One-time $49 — yours forever, Pro unlocks it too.

Customising

- Point the Postgres steps at your own tables and the agent queries follow.
- Adjust the fields in Set - Audit Trail to log the columns you need.