How to Automate Lead Qualification and Scoring With AI

The Lead Qualification Bottleneck

You get 50 leads per week from your website, social media, and referral partners. Each one fills out a form or sends a DM. You or your sales team spends 15 minutes per lead researching their company, checking their website, and deciding if they're worth a call.

That's 12.5 hours per week just on qualification. And the data shows that 70% of those leads will never buy — but you don't know which 70% until you've already wasted time on them.

AI lead scoring solves this. It analyses each lead in 3 seconds and ranks them by conversion probability. Your sales team only talks to the top 30%.

Manual vs AI Qualification

Lead Scoring Tools Compared

Best for African entrepreneurs: n8n + Groq. Free, fully customisable, and you own the scoring model.

Building the Scoring Model

Define Your Ideal Customer Profile

Before automating, document what makes a good lead for YOUR business:

Ideal Customer Profile (example for a Lagos SaaS agency):

Must-haves:
- Company size: 5–200 employees
- Revenue: ₦10M+ annually
- Location: Nigeria, Ghana, Kenya
- Has a website (not just social media)
- Decision-maker or influencer title

Disqualifiers:
- Student or job seeker
- Budget under ₦500K
- Outside Africa with no African operations

Step 1: Build the Lead Intake Workflow

Set up n8n to capture leads from multiple sources:

Sources:
  → Website form (Webhook)
  → LinkedIn DMs (via API)
  → WhatsApp Business (via Twilio)
  → Email inquiries (Gmail trigger)
  → Referral form (Airtable)
     ↓
  → Normalise data into standard format
  → Send to AI scoring node

Normalise Lead Data

// n8n Function Node — normalise incoming leads
const lead = $input.first().json;

return {
  json: {
    name: lead.name || lead.full_name || lead.Name || 'Unknown',
    email: lead.email || lead.Email || '',
    company: lead.company || lead.organization || lead.Company || '',
    title: lead.title || lead.role || lead.job_title || '',
    website: lead.website || lead.Website || lead.url || '',
    source: lead.source || 'unknown',
    message: lead.message || lead.body || lead.notes || '',
    phone: lead.phone || lead.Phone || '',
    country: lead.country || lead.Country || ''
  }
};

Step 2: AI Scoring Node

Send the normalised lead to Groq with a structured scoring prompt:

const scoringPrompt = `You are a lead qualification agent for a Nigerian SaaS agency.

Score this lead from 0-100 based on these criteria:

1. Company fit (0-25): Does the company match our ICP? (5-200 employees, ₦10M+ revenue)
2. Title/role fit (0-25): Is this person a decision-maker or influencer?
3. Budget signals (0-25): Any indicators of budget or purchasing intent?
4. Urgency signals (0-25): Any time-sensitive needs mentioned?

Lead data:
Name: ${lead.name}
Company: ${lead.company}
Title: ${lead.title}
Website: ${lead.website}
Message: ${lead.message}
Country: ${lead.country}
Source: ${lead.source}

Return ONLY valid JSON:
{
  "score": <0-100>,
  "tier": "hot" | "warm" | "cold",
  "reasoning": "<2 sentence explanation>",
  "suggested_action": "<specific next step>"
}`;

Step 3: Route Leads by Score

Based on the AI score, route leads automatically:

The Routing Logic

// n8n Switch Node logic
const score = aiResponse.score;

if (score >= 75) {
  // HOT: immediate response
  await sendSlackAlert(lead, score, aiResponse.reasoning);
  await sendPersonalizedEmail(lead, aiResponse.suggested_action);
  await createCRMTask(lead, 'Call within 1 hour', 'high');
} else if (score >= 50) {
  // WARM: nurture sequence
  await sendWarmNurtureEmail(lead);
  await addToCRMSequence(lead, 'warm-nurture');
} else if (score >= 25) {
  // COOL: light touch
  await sendNewsletterSignup(lead);
  await addToCRMSequence(lead, 'cool-nurture');
} else {
  // COLD: long-term only
  await addToLongTermNurture(lead);
}

Step 4: Track and Optimise Your Scoring Model: Track and Optimise Your Scoring Model

Log every lead's score, AI reasoning, and eventual outcome:

After 100+ leads, compare AI scores with actual outcomes. If leads scoring 60-70 convert at the same rate as leads scoring 80+, adjust your scoring criteria. The model improves with data.

Common Mistakes

Expected Results

A Lagos SaaS agency saw qualification time drop from 15 min/lead to 3 seconds, conversion on called leads jump from 12% to 28%, and revenue per sales hour triple — all for ₦0 in tools.

Your most expensive resource — sales time — gets pointed exclusively at leads most likely to buy. Everything else is handled automatically.