How to Build a Lead Qualification Bot Using Make.com and Claude API in 2026

This article mentions a tool I use; the link at the end is an affiliate link.

Lead qualification eats up hours that could be spent closing deals. In 2026, AI automation isn't just for enterprise—you can build a sophisticated lead qualification bot in an afternoon using Make.com and Claude API.

This guide walks you through creating a bot that reads inbound form submissions, asks intelligent follow-up questions via email, scores leads, and routes qualified prospects to your CRM—all without human intervention.

What You'll Build

A three-stage automation that:

1. Captures form submissions from your website
2. Uses Claude to analyze responses and generate personalized follow-up questions
3. Scores leads based on conversation quality and routes them appropriately

No coding required, though basic API familiarity helps.

Prerequisites

• Make.com account (free tier works for testing)
• Anthropic API key for Claude
• A form tool (Typeform, Google Forms, or custom HTML)
• Email service (Gmail, SendGrid, or Mailgun)
• Optional: CRM like Airtable or HubSpot

Step 1: Set Up Your Trigger Form

Create a simple contact form with these fields:

• Name
• Email
• Company
• Initial interest/question

In Make.com, create a new scenario and add a webhook trigger. Copy the webhook URL and configure your form to POST data there when submitted. Test it by submitting a dummy entry—you should see the data appear in Make.com.

Step 2: Connect Claude for Initial Analysis

Add an HTTP module in Make.com to call Claude API:

URL: https://api.anthropic.com/v1/messages

Headers:

• x-api-key: Your Anthropic API key
• anthropic-version: 2023-06-01
• content-type: application/json

Body (JSON):

{
  "model": "claude-3-5-sonnet-20241022",
  "max_tokens": 1024,
  "messages": [{
    "role": "user",
    "content": "Analyze this lead submission and generate 2-3 qualifying questions. Lead info: Name: {{name}}, Company: {{company}}, Interest: {{interest}}. Return ONLY a JSON object with: score (1-10), questions (array), reasoning."
  }]
}

Map the form fields to the prompt variables. Claude will return structured analysis you can parse in the next step.

Step 3: Parse and Route Based on Initial Score

Add a JSON parser module to extract Claude's response. Then add a Router with two paths:

Path A (Score 7+): High-potential leads
Path B (Score 1-6): Nurture leads

For Path A, send an immediate personalized email with Claude's generated questions. For Path B, add to a nurture sequence.

Step 4: Handle Email Responses

Create a second Make.com scenario triggered by incoming emails (use Gmail or email parser webhook).

When a lead replies:

1. Extract email body and thread history
2. Send to Claude again with context: "Previous questions: [questions]. Lead response: [response]. Should we qualify this lead? Return JSON with final_score, recommendation, next_action."
3. If final_score > 8, create a CRM entry and notify your sales team
4. If 5-8, send one more qualifying question
5. If < 5, add to long-term nurture

Step 5: Add CRM Integration

For qualified leads, add an Airtable or HubSpot module to create a new contact record. Map fields:

• Contact info from original form
• Conversation history
• AI qualification score
• Timestamp
• Source

This creates a clean handoff to sales with full context.

Real-World Optimization Tips

Prompt Engineering Matters: Your initial Claude prompt determines everything. Test with 20-30 real submissions and refine. I spent two weeks tweaking mine before the scoring became reliable.

Cost Control: Claude API calls cost roughly $0.003 per qualification. Set Make.com execution limits to avoid runaway costs during testing.

Response Time: Leads expect speed. Keep your scenarios under 30 seconds end-to-end by using efficient modules and avoiding unnecessary steps.

Human Override: Always include a manual review flag for edge cases. AI is good, not perfect.

Scaling Beyond the Basics

Once your bot runs smoothly, consider:

• Adding SMS notifications for hot leads
• Training Claude on your best customer profiles
• A/B testing different question sequences
• Integrating with calendar tools for automatic booking

When I was setting up the email nurture sequences, I used Perpetual Income 365 to structure the follow-up cadence and messaging templates, which saved time on the copywriting side. The framework helped organize the multi-touch approach, though you can absolutely build this manually with any email tool.

Common Pitfalls to Avoid

Over-automation: Don't remove humans entirely. Use AI to filter and prioritize, not replace judgment.

Ignoring GDPR/Privacy: Always include unsubscribe options and data handling notices.

Weak Prompts: Generic prompts produce generic results. Be specific about your ideal customer profile in the Claude instructions.

No Testing: Run 50+ test leads before going live. Edge cases will surprise you.

Measuring Success

Track these metrics:

• Qualification accuracy (compare AI scores to actual conversions)
• Time saved per lead
• Response rate to follow-up questions
• False positive rate (leads scored high but didn't convert)

Aim for 80%+ accuracy before trusting the system fully.

Next Steps

Start with the basic four-module flow: webhook → Claude analysis → email → CRM. Get that working, then add complexity.

The entire setup takes 3-4 hours for a working prototype. Refinement is ongoing, but you'll see time savings immediately.

This approach works for service businesses, SaaS, agencies, and info products. The key is customizing the qualification criteria to your specific business model.

---

The tool mentioned above is an affiliate link (disclosed at top): Perpetual Income 365