AutonomousHQ
← Tutorials
intermediate9 min read2026-04-07

How to Automate Lead Qualification with AI Tools

Build a fully automated lead qualification pipeline using AI tools that scores, enriches, and routes inbound leads without a sales team.

How to Automate Lead Qualification with AI Tools

Most solo founders and small teams spend hours each week manually reviewing inbound leads, copying data between tools, and deciding who is worth a follow-up. That process can be fully automated. This tutorial walks you through building a lead qualification pipeline that captures, scores, enriches, and routes leads automatically -- no sales team required.

By the end, you will have a system where a new form submission triggers an AI agent that scores the lead, pulls in company context from public sources, writes a personalised outreach draft, and routes high-value leads to your CRM or inbox.

What You Will Build

A pipeline with these stages:

  1. Lead captured via a web form
  2. Webhook fires to an automation platform
  3. AI agent scores the lead against your ideal customer profile
  4. Enrichment step pulls company size, industry, and tech stack
  5. High-score leads get a drafted outreach email
  6. Lead is written to your CRM with score and draft attached
  7. Low-score leads are tagged and archived

Tools used in this tutorial:

You can swap any of these for tools you already use. The logic is the same.

Step 1: Set Up Your Lead Capture Form

Create a form that collects the information your scoring model will need. At minimum:

  • Full name
  • Work email
  • Company name
  • What they are trying to solve (free text)
  • Current team size (dropdown)
  • Monthly budget range (optional but useful)

In Tally or Typeform, enable the webhook option and copy the webhook URL -- you will paste it into Make in the next step.

Keep the form short. Asking too many questions reduces completion rates. The enrichment step will fill in the gaps automatically.

Step 2: Create Your Make Scenario

In Make, create a new scenario and add a Webhook module as the trigger. Paste in the URL from your form tool and activate it.

Add the following modules in order:

  1. HTTP / Webhook -- receives form data
  2. HTTP: Make a Request -- calls the enrichment API
  3. OpenAI / Claude: Create Completion -- scores the lead
  4. Router -- branches based on score
  5. HubSpot / Notion: Create Record -- writes to CRM
  6. Gmail / SMTP: Send Email -- optional notification for high-score leads

Save the scenario but do not activate it yet. You will wire up each module in the steps below.

Step 3: Enrich the Lead

Before scoring, you want more context than the form provides. Use the enrichment module to pull in company data.

Using Clearbit Enrichment API:

In the HTTP module, set:

  • URL: https://person.clearbit.com/v2/combined/find?email={{email}}
  • Method: GET
  • Headers: Authorization: Bearer YOUR_CLEARBIT_KEY

Map the email field from the webhook payload to the email parameter.

The response will include company size, industry, estimated revenue, and technology stack. Map these fields to variables you will pass into the scoring prompt in the next step.

If you do not have a Clearbit account, Apollo.io offers a similar enrichment endpoint at a lower price point. The field names differ slightly but the logic is identical.

Step 4: Score the Lead with AI

This is the core of the pipeline. You will write a prompt that instructs the AI to evaluate the lead against your ideal customer profile and return a structured score.

In the OpenAI or Claude module, set up a completion with this system prompt:

You are a lead qualification assistant. Your job is to score inbound leads on a scale of 1 to 10 based on fit with our ideal customer profile.

Ideal customer profile:
- B2B SaaS or services company
- 10 to 200 employees
- Annual revenue between $1M and $50M
- Trying to reduce headcount or automate operations
- Based in US, UK, Canada, or Australia

Score criteria:
- 8 to 10: Strong fit. Route to priority outreach.
- 5 to 7: Moderate fit. Add to nurture sequence.
- 1 to 4: Poor fit. Archive.

Return a JSON object with this exact shape:
{
  "score": <number>,
  "reason": "<one sentence explaining the score>",
  "tier": "priority" | "nurture" | "archive"
}

Then set the user message to include the lead data you mapped from the form and enrichment step:

Name: {{name}}
Email: {{email}}
Company: {{company}}
Team size: {{team_size}}
Industry: {{industry}}
What they want: {{problem_statement}}
Budget: {{budget}}

Parse the JSON response. Make has a built-in JSON parse module you can use to extract score, reason, and tier as separate variables.

Step 5: Draft Outreach for High-Score Leads

For leads with tier = "priority", add a second AI call to draft an outreach email.

Use this system prompt:

You are an expert at writing short, direct outreach emails. Write a 3-sentence cold email for a founder reaching out to a prospect.

Rules:
- No flattery or filler phrases
- Reference something specific about their company or problem
- End with a single clear question, not a call to action
- Tone: direct, peer-to-peer, no corporate language
- Max 80 words

User message:

Prospect name: {{name}}
Company: {{company}}
Industry: {{industry}}
Their problem: {{problem_statement}}
Our product: [describe your product in one sentence here]

Store the drafted email as a variable. You will attach it to the CRM record in the next step.

Step 6: Write to Your CRM

For all leads, create a record in HubSpot, Notion, or your database of choice.

In HubSpot:

Use the HubSpot module and map:

  • Contact: name, email, company
  • Custom properties: score, tier, AI reasoning, drafted outreach email

In Notion:

Use the Notion module to create a page in your leads database. Set these properties:

  • Name (title)
  • Email
  • Score (number)
  • Tier (select: priority / nurture / archive)
  • Reason (text)
  • Draft email (text)
  • Source (text: "inbound form")
  • Date created (date: now)

This gives you a clean view of every lead with context attached, even if you never review the raw form submissions.

Step 7: Route and Notify

Add a Router module after the CRM write. Branch on the tier variable:

Priority branch:

  • Send yourself an email with the lead summary and drafted outreach
  • Optionally trigger a Slack or Discord notification

Nurture branch:

  • Add the contact to your email marketing tool (Mailchimp, ConvertKit, etc.) in a nurture sequence segment
  • No immediate notification required

Archive branch:

  • Write to CRM with archived: true
  • No further action

Step 8: Test the Full Flow

Before activating the scenario, run a manual test:

  1. Submit a fake lead through your form using a work email you control
  2. Check Make's execution log to verify each module received and passed data correctly
  3. Confirm the CRM record was created with the right fields
  4. If you used a priority lead, check that you received the notification email

Common issues to fix at this stage:

  • JSON parse errors if the AI returns extra text around the JSON object (fix by adjusting your prompt to say "return only valid JSON, no other text")
  • Enrichment returning empty fields for small or new companies (handle with fallback values in Make's mapping)
  • Rate limits on the AI API if you receive a burst of submissions (add an error handler and retry module)

Once the test passes, activate the scenario.

What to Measure

After your pipeline has been running for two weeks, review:

  • Score distribution: are most leads clustering at one end? Adjust your ICP definition.
  • AI accuracy: spot-check 10 priority leads. Did the score match your intuition?
  • Draft quality: are the outreach emails usable with minor edits, or do they need a full rewrite? Tune the prompt.
  • Conversion from drafted outreach: track reply rates on emails sent using the AI drafts versus emails you wrote from scratch.

Iterate the scoring prompt and ICP definition based on what you learn.

Next Steps

Once this pipeline is stable, you can extend it:

  • Connect LinkedIn outreach tools like Lemlist or Instantly to auto-send the drafted email for nurture-tier leads
  • Add a second enrichment pass using news search to flag leads that recently raised funding or hired aggressively
  • Build a weekly digest that summarises lead volume, average score, and tier breakdown, sent to your inbox automatically

The underlying pattern -- capture, enrich, score with AI, route, notify -- applies to any qualification problem. You can reuse this exact structure for job applications, partnership requests, or support tickets.