How to Build an AI Dealer Intelligence System That Scores, Filters & Auto-Messages Dealers — Anywhere in the World

01 The Problem: Sales Teams Flying Blind on Dealer Data

Picture this: your sales team has just pulled a list of 5,000 potential dealers from a marketplace. Hardware shops, fabricators, distributors, construction suppliers — all mixed together with no signal on which ones are actually worth calling. They start dialing randomly. Three hours later, they've reached 40 businesses, 30 of which were completely irrelevant.

This is the default state for most B2B manufacturers and distributors. Dealer acquisition is manual, slow, and expensive — not because the data doesn't exist, but because no one has built a system to prioritize it intelligently.

The AI dealer intelligence pipeline solves this entirely. Every lead gets a score. Every score maps to an action. Every action triggers automatically — without a human in the loop.

02 What This System Can Actually Do (Global Scope)

This is not a tool limited to one country or one marketplace. The pipeline is designed to work with any publicly listed dealer data source, anywhere in the world.

Supported Data Sources

Scale — No Upper Limit

• Firecrawl can process hundreds of thousands of pages per crawl job
• Google Sheets handles up to ~500,000 rows comfortably
• For larger datasets, swap Sheets for BigQuery or Supabase — same scoring logic applies
• Twilio WhatsApp scales to millions of messages per month on a business account

03 System Architecture Overview

Five components, all free or near-zero cost at the core level:

04 The AI Dealer Scoring Engine (1–10 Logic)

Every dealer gets a score from 1 to 10 based on three weighted signals: business type fit, location tier, and verification status. The score determines the bucket — and the bucket determines the automated action.

Base Score by Business Type

Customize the categories and weights to match your product. The example below is for a building materials manufacturer:

Score Modifiers

The Four Dealer Buckets

05 Step 1 — Scrape Dealer Data from Any Marketplace

Firecrawl is a web crawling API that returns structured data from any website. Point it at your marketplace search results and it returns clean JSON or CSV — no manual copy-paste needed.

// Example search URL patterns (replace with your marketplace + product)
// IndiaMart (India)
https://www.indiamart.com/search.mp?ss=YOUR+PRODUCT+KEYWORD

// Thomasnet (USA)
https://www.thomasnet.com/search/?searchterm=YOUR+PRODUCT

// Alibaba (Global)
https://www.alibaba.com/trade/search?SearchText=YOUR+PRODUCT

// Google Maps (any country)
https://www.google.com/maps/search/dealer+YOUR+PRODUCT+CITY
URL patterns

Your scraped output must have at minimum these columns for the scoring engine to work:

Company, City, State/Region, Phone, Source URL, Verified (GST/VAT/Tax)

// Example rows (fictional companies for illustration):
"RoofTech Supplies Ltd",   "Example City", "Region A", "98XXXXXXXX", "https://...", "Verified"
"AlphaFab Works",           "Example City", "Region B", "70XXXXXXXX", "https://...", "Not Verified"
"Metro Hardware Depot",     "Metro City",   "Region A", "91XXXXXXXX", "https://...", "Verified"
CSV

06 Step 2 — Score & Filter with Google Apps Script

The scoring logic runs inside Google Apps Script — a free JavaScript environment attached to your Google Sheet. Here's the complete, adaptable scoring function:

function scoreDealer(businessType, city, verified, competitors) {
  let score = 5.0; // default baseline for unknown types

  // --- Step 1: Business type base score ---
  const typeMap = {
    'roofing':     9.5,  'sheet':        9.5,
    'aluminium':   9.0,  'glazing':      9.0,
    'hardware':    8.5,  'timber':       8.5,
    'fabrication': 8.5,  'steel':        8.5,
    'plastic':     8.0,  'polymer':      8.0,
    'engineering': 7.0,  'construction': 7.0
  };

  const lowerType = businessType.toLowerCase();
  for (const [keyword, baseScore] of Object.entries(typeMap)) {
    if (lowerType.includes(keyword)) { score = baseScore; break; }
  }

  // --- Step 2: Auto-remove competitors ---
  const lowerComp = competitors.map(c => c.toLowerCase());
  if (lowerComp.some(c => lowerType.includes(c))) {
    return { score: 0, bucket: 'IGNORE', flag: 'COMPETITOR' };
  }

  // --- Step 3: Verification modifier ---
  if (verified === 'Verified') score += 0.5;

  // --- Step 4: City tier modifier ---
  const tier1Cities = ['metro', 'capital', 'port'];
  const tier2Cities = ['secondary', 'regional'];
  const lowerCity = city.toLowerCase();
  if (tier1Cities.some(c => lowerCity.includes(c))) score += 0.3;
  else if (tier2Cities.some(c => lowerCity.includes(c))) score += 0.1;

  score = Math.min(score, 10);

  // --- Step 5: Assign bucket ---
  const bucket =
    score >= 9.0 ? 'HOT AUTO LEAD' :
    score >= 7.5 ? 'VERIFY FIRST'  :
    score >= 5.0 ? 'NURTURE'        : 'IGNORE';

  return { score, bucket, flag: 'OK' };
}
Google Apps Script

Apply Color-Coding to Rows

function colorByBucket(sheet, row, bucket) {
  const palette = {
    'HOT AUTO LEAD': '#d9ead3',
    'VERIFY FIRST':  '#fce5cd',
    'NURTURE':       '#cfe2f3',
    'IGNORE':        '#f4cccc'
  };
  sheet.getRange(row, 1, 1, 13).setBackground(palette[bucket] || '#ffffff');
}
Google Apps Script

07 Step 3 — Build the 5-Tab Live Dashboard

Organize the Google Sheet into five purpose-built tabs, each serving a different function in your sales workflow:

Column Structure (All Dealers Tab)

A: Company Name    B: City          C: State / Region
D: Business Type   E: Phone         F: Category
G: Verified        H: Source        I: Score
J: Bucket          K: WhatsApp Draft
L: Next Action     M: Flag
Sheet Structure

08 Step 4 — Automate WhatsApp Outreach with Twilio

Once dealers are scored and bucketed, the system fires personalized WhatsApp messages via Twilio's WhatsApp API. Two flows run automatically: outreach to dealers, and internal alerts to your sales reps.

Setup: Twilio WhatsApp Sandbox (Free Testing)

1. Create a free account at twilio.com
2. Go to Messaging → Try it out → Send a WhatsApp message
3. Note your sandbox number and join keyword
4. Send the join keyword from your test phone to activate
5. Copy your Account SID and Auth Token from the Twilio Console dashboard

Dealer Outreach Message Function

function sendDealerOutreach(phone, companyName, city, productCategory) {
  const ACCOUNT_SID = 'YOUR_TWILIO_ACCOUNT_SID';
  const AUTH_TOKEN  = 'YOUR_TWILIO_AUTH_TOKEN';
  const FROM_NUMBER = 'whatsapp:+1XXXXXXXXXX';
  const TO_NUMBER   = \`whatsapp:+\${phone}\`;

  const messageBody = [
    \`Hello, we are reaching out to \${companyName} in \${city}.\`,
    \`We manufacture \${productCategory} and are looking for authorized\`,
    \`dealers in your area. Interested in a partnership?\`,
    \`Reply YES for product details and pricing.\`
  ].join(' ');

  const endpoint = \`https://api.twilio.com/2010-04-01/Accounts/\${ACCOUNT_SID}/Messages.json\`;
  const credentials = Utilities.base64Encode(\`\${ACCOUNT_SID}:\${AUTH_TOKEN}\`);

  UrlFetchApp.fetch(endpoint, {
    method:  'post',
    headers: { Authorization: \`Basic \${credentials}\` },
    payload: { From: FROM_NUMBER, To: TO_NUMBER, Body: messageBody }
  });
}
Google Apps Script

Internal Sales Rep Alert Function

function sendRepAlert(repPhone, dealer) {
  const alertLines = [
    \`HOT LEAD ALERT\`,
    \`Company : \${dealer.company}\`,
    \`Score   : \${dealer.score}/10\`,
    \`Phone   : \${dealer.phone}\`,
    \`City    : \${dealer.city}\`,
    \`Bucket  : \${dealer.bucket}\`,
    \`Action  : Call within 24 hours\`
  ];
  // use same Twilio fetch with TO = repPhone
}
Google Apps Script

09 Step 5 — Deploy, Trigger & Run on Autopilot

Open your Google Sheet → Extensions → Apps Script. Paste in all functions (scoring, bucketing, formatting, Twilio). Save and run once to grant the required permissions: Sheets, UrlFetch, Drive.

function runDealerIntelligence() {
  clearAllSheets();                   // wipe old data
  const raw    = loadDealersFromCSV(); // read uploaded CSV
  const scored = scoreDealers(raw);    // apply scoring engine
  writeToAllSheets(scored);            // populate all 5 tabs
  applyBucketFormatting();             // color-code rows
  sendHotLeadAlerts(scored);          // fire WhatsApp alerts
}
Google Apps Script

In Apps Script → Triggers, add a time-driven trigger: run runDealerIntelligence every morning at 8 AM. New data gets processed overnight. Your team wakes up to a fully scored, color-coded, WhatsApp-ready lead dashboard — every single day.

10 Sample Output: Scored Dealer Table

Below is an example output table using fictional company names to illustrate what the scored dashboard looks like in practice.

All company names above are fictional examples only. Competitor dealers are auto-flagged and removed before this table is generated.

11 Scaling Beyond Google Sheets

Google Sheets is the right starting point — zero cost, instant setup, no infrastructure. But as your pipeline grows, here's how to scale each layer:

The scoring logic — the JavaScript functions — remains identical at every scale level. Only the data storage and trigger layer changes. Abstract your business logic from your infrastructure from day one.

12 What's Next: Make.com, AI Personalization & CRM Sync

This guide is the foundation. Each of these layers can be added incrementally — start with the sheet, add Make.com when your volume grows, and layer in AI personalization when your messaging needs to feel more human.

References & Further Reading

• Google Apps Script Documentation — Official guide to automating Google Sheets and Workspace
• Twilio WhatsApp API Docs — Complete reference for sending WhatsApp messages programmatically
• n8n Documentation — How to build no-code automation workflows and connect business tools
• Google BigQuery Docs — Scaling data pipelines from Sheets to enterprise-grade analytics
• Wikipedia: Lead Scoring — What it is and how AI improves B2B lead qualification