What this template does

Every new lead in your Airtable lead capture base triggers an automated qualification flow. The scenario sends the lead's data to GPT-4 along with your ICP definition, GPT-4 returns a structured score (0-100) and reasoning. Based on the score, leads are automatically routed to the right destination.

This eliminates the manual triage step that typically eats 30-45 minutes per day for a sales team.

Use cases

  • B2B SaaS qualifying inbound demo requests
  • Agencies filtering inbound contact forms
  • E-commerce brands routing wholesale inquiries
  • Coaches pre-qualifying discovery call bookings

How it works

  1. Watch new Airtable rows

    The trigger module watches your Lead Capture base for new rows added in the last 5 minutes.

  2. Build the qualification prompt

    Use a Set Variable module to construct the GPT-4 prompt with your ICP definition + the new lead's data.

  3. Call GPT-4 with structured output

    Send to OpenAI with response_format: json_object. Ask for a JSON with score, reasoning, recommended_action.

  4. Parse JSON response

    Use the JSON Parse module to extract the structured fields from GPT-4's response.

  5. Update Airtable with score

    Write the score, reasoning, and recommended_action back to the lead's row.

  6. Router based on score

    If score >= 80, route to sales. If 50-79, route to nurture. If < 50, mark as disqualified.

  7. Notify Slack for high-fit leads

    Send a formatted message to #sales-alerts with the lead's name, company, and score.

  8. Add to nurture sequence

    For medium-fit leads, add to your Mailchimp/Brevo nurture list with a tag for the recommended segment.

Setup time

Around 15 minutes if you already have Airtable, OpenAI, and Slack connected to Make.

Customization ideas

  • Add multiple AI models for cross-validation (GPT-4 + Claude) and average the scores
  • Include LinkedIn enrichment via Clearbit before the AI scoring step
  • Adapt the scoring rubric per product line (different ICPs for different products)
  • Add a feedback loop: when sales close a deal, log the original score for accuracy tracking