A complete lead pipeline: inquiry form submission → AI qualification scoring → branch by tier → AI-drafted personalized reply → CRM entry → Slack team alert. Zero manual steps.
Design Before You Build
Complex automations built without a plan become unmaintainable spaghetti. Before opening Make.com, answer four questions on paper:
1. What is the trigger?
→ Form submission via Typeform/Google Forms
2. What is the happy path?
→ Qualified lead gets personalized email + enters CRM
3. What can go wrong?
→ Spam submission, API failure, malformed JSON
4. What does success look like?
→ Lead in CRM + team notified + reply sent within 5 minDrawing a quick flowchart before building saves hours of backtracking. The whole thing takes 10 minutes and prevents building into a dead end.
AI Lead Qualification
The highest-value thing AI can do in a lead system is qualify leads consistently. You define the criteria once; AI applies them to every submission the same way every time — no fatigue, no bias based on whether it's your first or fiftieth inquiry of the day.
Evaluate this business inquiry. Return JSON only.
Criteria:
- Budget signal: Do they suggest budget over $5K?
- Timeline: Starting within 3 months?
- Fit: Does their need match AI training services?
{
"score": 1-10,
"tier": "hot | warm | cold",
"reasoning": "1 sentence",
"reply_urgency": "immediate | same-day | this-week"
}
Inquiry: {{inquiry}}The score + tier combination lets you route the workflow without ambiguity. Your router conditions become simple: tier = "hot", tier = "warm", tier = "cold".
Build the Complete Pipeline
Here's the full scenario, module by module:
Module 1: Trigger. Google Forms or Typeform — Watch New Submissions. Your contact form feeds this.
Module 2: AI Qualification. OpenAI — score the lead. Parse the JSON output.
Module 3: Router. Add a Router (the diamond icon). Create three filter conditions:
- Branch A:
tier = "hot" - Branch B:
tier = "warm" - Branch C: else (cold)
Branch A (Hot): Fast reply + priority CRM + instant Slack alert
Add an OpenAI step: "Write a warm, specific reply to this inquiry. Reference what they said. Keep it under 130 words. Close with a calendar link offer." Pass the original inquiry + your services description.
Gmail — Send Email with the AI-drafted reply. CC yourself.
CRM entry (Notion/Airtable/HubSpot) with tag: Priority.
Slack — post to #leads: "🔥 HOT LEAD — [name] — Score [score] — [reasoning]"
Branch B (Warm): Standard reply + CRM
Similar AI reply step but prompt says "professional and warm, less urgent tone."
CRM entry with tag: Follow-up.
Slack — daily digest mention (not immediate alert).
Branch C (Cold): CRM only
CRM entry with tag: Low Priority. No reply sent, no notification.
Run this for 30 days, then look at which "hot" leads actually converted. Did the AI qualification match reality? Adjust the criteria in your prompt based on what you learn. This is how you improve an automation over time — not by rebuilding it, but by refining the prompt logic.
What You Learned Today
- How to design a complete automation system before building it
- How to use AI for consistent, scalable lead qualification
- How to branch Make.com workflows with Routers based on AI output
- How to chain AI qualification + personalized reply generation
Go Further on Your Own
- Add a spam filter: check if the email domain is in a known spam list (you maintain in a Google Sheet) and route those to a 'Review' folder without sending any reply
- Build a weekly report: every Monday, pull the previous week's CRM entries and have AI summarize trends — what industries, what budgets, what pain points appeared most
- Test your qualification by submitting 10 fictional inquiries ranging from obvious spam to obvious hot leads — verify the AI scores each correctly, then adjust the prompt if it doesn't
Course Complete!
You finished all 5 days. Ready to go deeper?
Reserve Your Bootcamp SeatWant live instruction and hands-on projects? Join the AI bootcamp — 3 days, 5 cities.