Build a Client Intake Chatbot with n8n and OpenAI in Under an Hour

If you're running a service business, your intake process is leaking money. Let's fix it with n8n, OpenAI, and about 45 minutes of setup.

What We're Building

A webhook-triggered workflow that:

1. Receives a user message from a simple HTML form
2. Sends it to GPT-4o-mini with a system prompt that qualifies the lead
3. Stores the conversation in Airtable
4. Posts a summary to a Slack channel if the lead looks qualified

No custom server. No code deployments. Just nodes.

Prerequisites

• n8n Cloud account (free tier works) or self-hosted instance
• OpenAI API key
• Airtable account with a base set up (I'll show the schema)
• Slack webhook URL

Step 1: The Trigger Node

Create a new workflow and add a Webhook trigger node. Set the HTTP method to POST and copy the generated URL — you'll drop this into your form's action attribute.

{
  "event": "intake_message",
  "user_id": "{{$json.user_id}}",
  "message": "{{$json.message}}"
}

Step 2: The OpenAI Node

Add an OpenAI node after the webhook. Use the Chat Completion operation with gpt-4o-mini. Set your system prompt to something like:

You are a lead qualification assistant for a web development agency. Ask clarifying questions about project scope, budget, and timeline. After 3-4 exchanges, output a JSON object with keys: qualified (boolean), budget_range, project_type, summary.

Step 3: Airtable + Slack

Wire the response into an Airtable node to append a row, then add a conditional branch — if qualified === true, fire the Slack notification.

Wrap-Up

The whole thing runs serverlessly and costs pennies per conversation. The real power is the system prompt — spend most of your time there.