How to Build a WhatsApp Bot for Your Business: Step-by-Step Guide (2026)

Building a WhatsApp bot requires three components: WhatsApp Business API access (from ILS 350/month via a BSP like 360dialog), an automation platform like N8N to handle message logic, and optionally an AI model (GPT-4o or Claude) for intelligent responses. The complete process takes 1-4 weeks depending on complexity. This guide walks through every step — from API setup to CRM integration — with specific instructions for businesses in Israel, including Hebrew bot configuration and WhatsApp's 24-hour messaging rules.

What You Need Before Starting

Prerequisites Checklist

Choosing Your Bot Type

Before building, decide which type of bot matches your needs:

Build cost follows those drivers, not the label. See the pricing page for how an engagement is structured.

Step 1: Set Up WhatsApp Business API (Days 1-3)

1.1 Create and Verify Meta Business Account

1. Navigate to business.facebook.com
2. Create a Business Account (or log into existing)
3. Go to Security Center → Start Verification
4. Provide:
   • Israeli business registration number (ח.פ. or ע.מ.)
   • Business address
   • Business phone number
   • Business website (verified via DNS)
5. Upload supporting document (business registration certificate)

Timeline: Meta typically verifies within 1-3 business days.

1.2 Choose and Set Up a BSP

For this guide, we use 360dialog (cost-effective, API-first approach):

1. Sign up at 360dialog.com
2. Connect your Meta Business Account
3. Register your phone number for WhatsApp API
4. Complete phone number verification (SMS or voice call)
5. Receive API key and webhook URL

Alternative BSPs for Israeli businesses:

• WATI — Best for non-technical teams (dashboard interface)
• Twilio — Best if you already use Twilio for SMS
• Infobip — Best for enterprise with local support needs

1.3 Create Your First Message Templates

Submit at least 3 templates to Meta for approval:

Template 1: Welcome/Lead Capture

Name: welcome_message
Language: Hebrew
Category: Marketing
Body: "שלום {{1}}! תודה שפנית ל-{{2}}.
איך נוכל לעזור לך היום?
1️⃣ מידע על שירותים
2️⃣ קביעת פגישה
3️⃣ שיחה עם נציג"

Template 2: Appointment Reminder

Name: appointment_reminder
Language: Hebrew
Category: Utility
Body: "שלום {{1}}, תזכורת: יש לך פגישה מחר
בתאריך {{2}} בשעה {{3}}.
לאישור — שלח/י 1
לשינוי — שלח/י 2"

Template 3: Follow-Up

Name: follow_up
Language: Hebrew
Category: Marketing
Body: "שלום {{1}}, שמנו לב שהתעניינת ב{{2}}.
רוצה לשמוע עוד? צוות המומחים שלנו ישמח לעזור.
📞 או שלח/י הודעה כאן"

Approval timeline: 24-48 hours per template.

Step 2: Set Up N8N (Day 2-3)

2.1 Install N8N

Option A: Self-Hosted (Recommended for cost control)

Using Docker on a VPS (DigitalOcean, Hetzner, AWS):

docker run -d --name n8n \
  -p 5678:5678 \
  -v n8n_data:/home/node/.n8n \
  -e N8N_BASIC_AUTH_ACTIVE=true \
  -e N8N_BASIC_AUTH_USER=admin \
  -e N8N_BASIC_AUTH_PASSWORD=your-secure-password \
  n8nio/n8n

Estimated hosting cost: ILS 80-200/month for a basic VPS.

Option B: N8N Cloud

Sign up at n8n.cloud. Plans start at EUR 24/month (2,500 executions).

2.2 Configure WhatsApp Webhook in N8N

1. Create a new workflow in N8N
2. Add a Webhook node as the trigger
   • Method: POST
   • Path: /whatsapp-incoming
   • Copy the webhook URL (e.g., https://your-n8n-domain.com/webhook/whatsapp-incoming)
3. Configure your BSP (360dialog) to send incoming messages to this webhook URL
4. Test by sending a WhatsApp message to your registered number

The webhook receives JSON data like:

{
  "messages": [{
    "from": "972501234567",
    "type": "text",
    "text": { "body": "שלום, מתעניין בשירותים שלכם" },
    "timestamp": "1708700000"
  }]
}

Step 3: Build Your Bot Logic (Days 3-7)

3.1 Basic Auto-Responder (No AI)

This is the simplest bot — responds with predefined messages based on keywords.

N8N Workflow:

1. Webhook Node — Receives incoming messages
2. Set Node — Extracts message text and sender number
3. Switch Node — Routes based on message content:
   • Contains "מחיר" or "price" → Send pricing info
   • Contains "פגישה" or "meeting" → Send booking link
   • Contains "שירות" or "service" → Send service list
   • Default → Send general greeting
4. HTTP Request Node — Sends response via WhatsApp API

Response example (pricing inquiry):

{
  "messaging_product": "whatsapp",
  "to": "{{sender_number}}",
  "type": "text",
  "text": {
    "body": "אין אצלנו מחירון לפי שירות — המחיר נגזר מהמערכות, מהנתונים ומהחריגים. רוצה לקבוע שיחת בירור?"
  }
}

3.2 AI-Powered Bot (Using GPT-4o)

For intelligent, context-aware conversations:

N8N Workflow:

1. Webhook Node — Receives incoming messages
2. Set Node — Extracts message text and sender number
3. AI Agent Node — Processes message with GPT-4o:
   • System Prompt: Business-specific instructions in Hebrew
   • Tools: CRM lookup, appointment booking, FAQ search
   • Memory: Conversation history per phone number
4. HTTP Request Node — Sends AI-generated response via WhatsApp API
5. Zoho CRM Node — Creates/updates lead record

System prompt example:

You are a business assistant for [Company Name], an automation company in Israel.
Your role: Qualify leads and schedule consultations.

Rules:
- Respond in Hebrew unless the customer writes in English
- Be direct and professional (Israeli business style)
- Ask about: business type, team size, main challenge, budget range
- After qualification, offer to schedule a free consultation
- Never make up pricing or capabilities
- If you don't know something, say you'll connect them with a team member

Services and pricing:
- Free initial call; the characterization phase itself is a paid, fixed-fee stage

Working hours: Sunday-Thursday 9:00-18:00

3.3 Interactive Menu Bot (Buttons and Lists)

WhatsApp supports interactive message types that work well for menu-driven bots:

Interactive Button Message:

{
  "messaging_product": "whatsapp",
  "to": "{{sender_number}}",
  "type": "interactive",
  "interactive": {
    "type": "button",
    "body": {
      "text": "שלום! איך אוכל לעזור?"
    },
    "action": {
      "buttons": [
        { "type": "reply", "reply": { "id": "services", "title": "שירותים" }},
        { "type": "reply", "reply": { "id": "pricing", "title": "מחירון" }},
        { "type": "reply", "reply": { "id": "meeting", "title": "קביעת פגישה" }}
      ]
    }
  }
}

Interactive List Message:

{
  "messaging_product": "whatsapp",
  "type": "interactive",
  "interactive": {
    "type": "list",
    "body": { "text": "בחר/י שירות:" },
    "action": {
      "button": "הצג שירותים",
      "sections": [{
        "title": "השירותים שלנו",
        "rows": [
          { "id": "whatsapp-bot", "title": "בוט לוואטסאפ", "description": "סוכן וואטסאפ מחובר למערכות" },
          { "id": "crm", "title": "הטמעת CRM", "description": "מערכת ניהול מותאמת או חיבור לקיימת" },
          { "id": "ai-agent", "title": "סוכן AI", "description": "סוכן שמבצע פעולות במערכות" },
          { "id": "automation", "title": "אוטומציה עסקית", "description": "אוטומציית תהליכים בין מערכות" }
        ]
      }]
    }
  }
}

Step 4: Connect to CRM (Week 2)

4.1 Zoho CRM Integration via N8N

Add CRM nodes to your N8N workflow to automatically manage leads:

Lead Capture Flow:

1. After AI agent qualifies the lead, extract: name, phone, interest, budget
2. Use N8N's Zoho CRM Node → Create Lead:
   • Lead Source: "WhatsApp Bot"
   • Phone: Customer's WhatsApp number
   • Description: AI agent's conversation summary
   • Custom fields: Budget range, timeline, specific interest
3. Set up Zoho CRM assignment rules to route leads to the right sales person

4.2 Conversation Logging

Log every WhatsApp conversation in CRM for sales team reference:

1. After each conversation turn, use N8N's Zoho CRM Node → Create Note
2. Attach note to the lead/contact record
3. Include: timestamp, customer message, bot response, any actions taken

This gives sales representatives full context when they follow up with qualified leads.

4.3 Automated Follow-Up Sequences

Build N8N scheduled workflows that:

• Day 1: Send WhatsApp thank-you message after conversation
• Day 3: Follow up if no response from qualified leads
• Day 7: Send relevant content (blog post, case study) via WhatsApp template
• Day 14: Final follow-up with special offer

Step 5: Test Your Bot (Days 10-14)

Testing Checklist

Common Issues and Fixes

Step 6: Launch and Monitor (Week 3+)

Soft Launch Strategy

1. Week 1: Route 25% of incoming WhatsApp messages to bot, rest to human agents
2. Week 2: Increase to 50% if quality metrics are acceptable
3. Week 3: Full rollout (100%) with human escalation for complex queries

Monitoring Dashboard

Track these metrics daily using N8N's built-in logging or a dashboard tool:

How much does it cost?

A complete starting point — a WhatsApp AI agent including the business number, server and agent control room — is from ILS 10,200; an agent with a management system (CRM) from ILS 18,200, a multi-agent system from ILS 25,350. There is no monthly fee; third-party costs (server, messages, AI) of about ILS 100–500 a month are paid directly to the providers. The full component price list: pricing page.

Option 1: DIY with N8N (Technical Team Available)

Option 2: Professional Implementation

How to Judge It Without Assuming a Payback Period

Do not start from a payback estimate; start from a measured baseline. Over one defined period, capture:

• hours actually spent on WhatsApp conversations, at your own loaded hourly cost;
• inquiries that arrived outside working hours and went unanswered;
• exceptions and rework the current process produces.

Then set the success thresholds before the build — accuracy, handling time and cost per action — and compare them against real post-launch measurement. If the net benefit is not positive or cannot be measured, there is no basis for a payback claim. We do not publish a typical payback period, because a return that was not measured in your own process is not evidence.

Advanced: Adding AI Agent Capabilities

Once your basic WhatsApp bot works, upgrade it with AI agent features:

Feature 1: Conversation Memory

Store conversation context in a database so the bot remembers previous interactions:

• N8N stores conversation history keyed by phone number
• AI agent retrieves history before each response
• Enables: "Last time you asked about CRM pricing. Have you made a decision?"

Feature 2: Product/Service Lookup

Connect the bot to your product database:

• N8N queries your database based on customer questions
• AI agent formats relevant product information
• Enables: "We have 3 packages that fit your budget. Here is the comparison..."

Feature 3: Appointment Booking

Integrate with Google Calendar or Calendly:

• AI agent checks available time slots
• Customer selects preferred time via WhatsApp interactive list
• N8N books the appointment and sends confirmation
• Automated reminders at 24 hours and 2 hours before

Feature 4: Multi-Language Support

Configure the AI agent to detect and respond in the customer's language:

• System prompt instructs: "Detect the customer's language and respond in the same language"
• Supports Hebrew, English, Arabic, and Russian (common in Israeli market)
• Language detection is automatic with GPT-4o and Claude

Frequently Asked Questions

Can I build a WhatsApp bot with Zapier instead of N8N?

Zapier can send WhatsApp messages through some integrations, but it lacks webhook support for receiving messages in real-time. Zapier is not recommended for WhatsApp bot development. N8N and Make are better suited because they support webhooks, branching logic, and AI nodes needed for conversational bots.

What happens if WhatsApp blocks my number?

Meta blocks numbers with poor quality ratings (high spam reports, low opt-in compliance). Prevention: (1) only message contacts who opted in, (2) include opt-out options, (3) send relevant, valuable messages. If blocked, contact your BSP for appeal — the process takes 1-2 weeks.

Can the bot transfer conversations to a human agent?

Yes. Configure an escalation trigger in N8N: when the customer types "agent" or "representative," or when the AI agent detects it cannot handle the request, the workflow notifies a human agent (via Slack, email, or WhatsApp group) with the conversation context. The human takes over within the same WhatsApp thread.

How many concurrent conversations can the bot handle?

N8N and WhatsApp API can handle hundreds of concurrent conversations without performance issues. The bottleneck is usually the LLM API rate limit: GPT-4o supports 10,000+ requests/minute on paid plans. For most Israeli SMBs, concurrent capacity is never a practical concern.

Can I add the bot to a WhatsApp Group?

WhatsApp Business API does not support group messaging. The bot operates in individual (1-to-1) conversations only. For group communication needs, consider integrating with Slack or Microsoft Teams via N8N.

Summary

Building a WhatsApp bot for your business involves 6 steps: set up WhatsApp Business API through a BSP, install N8N as your automation platform, build bot logic (keyword-based or AI-powered), connect to your CRM for lead management, test thoroughly, and launch with monitoring.

Running it yourself costs from about ILS 580/month in third-party fees. Whether it pays back is not something to assume in advance — it is measured against the success thresholds agreed before the build.

Need help building your WhatsApp bot? Automaziot AI specializes in WhatsApp bot development with AI capabilities, CRM integration, and Hebrew language support. Book a free consultation to discuss your project.

Last updated: August 2026