WhatsApp AI Bot India : Complete Setup Guide for Meta Business API in 2026

Key takeaway: Building a WhatsApp AI bot in India is not a chatbot plugin install. It requires Meta Business API access, a verified Business Manager, an approved phone number, and an automation layer that connects conversations to your CRM. This guide walks through each step in order.

• Step 1: Create and verify Meta Business Manager
• Step 2: Register a WhatsApp Business Account
• Step 3: Choose your automation layer
• Step 4: Build your first conversation flow
• Step 5: Integrate with CRM
• Pricing: conversation-based billing explained
• Common mistakes and how to avoid them
• Frequently asked questions

What the Meta Business API is

The WhatsApp consumer app and WhatsApp Business app are not the same as the WhatsApp Business Platform (Cloud API). Production AI bots in India run on the Business Platform, which provides:

• Programmatic send and receive via API
• Template messages for outbound notifications outside the 24-hour session window
• Webhook events for inbound messages
• Official business profile and green check verification path

You cannot scale a real AI bot on the free Business app alone. You need API access through Meta or an approved Business Solution Provider (BSP).

Our WhatsApp AI bot implementations always start with API readiness, not prompt design.

Step 1: Create and verify Meta Business Manager

Create Business Manager

Go to business.facebook.com and create a Business Manager account with your legal business name matching your GST registration or incorporation documents. Mismatched names delay verification.

Add business details

• Legal entity name and address
• Business website with privacy policy and contact page
• Business email on your domain (not a free Gmail address for verification)

Start business verification

Meta reviews documents to confirm the business is real. For Indian companies, typical documents include:

• Certificate of incorporation or GST certificate
• Utility bill or bank statement showing business address
• Business website live with product or service description

Verification can take a few days to several weeks. Start this before you build conversation logic.

Assign roles

Give your implementation partner or internal admin Developer and Business Admin roles as needed. Do not share personal Facebook passwords; use role-based access.

Step 2: Register a WhatsApp Business Account

Add WhatsApp product

Inside Business Manager, add the WhatsApp product and create a WhatsApp Business Account (WABA).

Phone number selection

You can use:

• A new number dedicated to API use (recommended)
• Migrate an existing number from the Business app (requires careful migration steps; the number leaves the app)

The number must receive SMS or voice OTP for verification. VoIP numbers often fail verification.

Display name approval

Your public WhatsApp business display name must match your brand and Meta commerce policies. Rejected names block go-live.

Connect Cloud API

Meta Cloud API hosts the webhook endpoint configuration. Register callback URL and verify token for inbound message events.

Message templates

Outbound messages outside an active 24-hour customer service window require pre-approved templates. Plan templates early for:

• Order confirmations
• Appointment reminders
• Payment links
• OTP or verification codes where permitted

Template approval adds lead time. Factor it into launch schedules.

Step 3: Choose your automation layer

The API delivers messages. Something else must decide replies. Common choices for Indian businesses:

Layer	Best for	India notes
n8n	Custom flows, CRM integration, voice handoff	Self-hosted or cloud; cost-effective at scale
Make	Visual automation, mid-complexity	Per-operation pricing adds up at volume
Zapier	Simple triggers, fast prototypes	Higher per-task cost for high message volume
Custom backend	Full control, large engineering team	Highest build and maintenance burden

We recommend n8n for most Indian SMEs connecting WhatsApp to HubSpot, Zoho, voice callbacks on Retell or Vapi, and internal databases. See our AI automation approach for hub-and-spoke design.

Voice plus WhatsApp is a common pattern: AI voice call for urgent qualification, WhatsApp for documents, links, and async follow-up.

Step 4: Build your first conversation flow

Start with one high-volume intent, not a general-purpose bot.

Example: lead capture after ad click

1. User clicks Click-to-WhatsApp ad or scans QR
2. Welcome message within session window asks two qualification questions
3. n8n writes responses to CRM and assigns owner
4. If qualified, trigger outbound voice callback via Retell or Vapi within 60 seconds
5. If unqualified, send polite closure message and tag CRM

AI reply layer

Connect an LLM with strict system prompt boundaries:

• Approved product facts only
• No invented pricing or policy promises
• Escalation keyword detection for human handoff
• Language match: reply in Hindi, English, or Hinglish as the user writes

Session window rules

Free-form AI replies are allowed inside the 24-hour customer service window after user message. Outside that window, use approved templates only.

Log every inbound and outbound message ID for billing reconciliation and support debugging.

Step 5: Integrate with CRM

WhatsApp without CRM sync creates another silo. Minimum integration requirements:

• Create or update contact on first message
• Append conversation summary to timeline
• Set lead source to WhatsApp campaign or ad ID
• Trigger tasks for human agents on escalation keywords
• Sync opt-out preferences for DPDPA compliance

HubSpot, Zoho CRM, and Salesforce all integrate via n8n. Our CRM integration patterns apply directly to WhatsApp events.

Bidirectional sync prevents automation from overwriting manual rep updates without rules.

Pricing: conversation-based billing explained

WhatsApp Business Platform billing is conversation-based, not per-message in the traditional SMS sense. Meta groups messages into 24-hour conversation windows by category:

• Marketing conversations: promotional outbound (template-based)
• Utility conversations: transactional updates (order status, appointment confirm)
• Authentication conversations: OTP and verification flows
• Service conversations: user-initiated support within the service window

Rates vary by country and category. India pricing is published on Meta's rate cards and changes periodically. Budget for:

• Meta conversation fees
• BSP markup if you use a provider instead of direct Cloud API
• LLM inference for AI replies
• n8n hosting or cloud execution costs

A bot that blasts marketing templates without session discipline gets expensive quickly. Design flows to maximise service and utility categories where appropriate.

Common mistakes and how to avoid them

Mistake 1: Starting with the chatbot UI before verification
Fix: Complete Business Manager verification and number registration first.

Mistake 2: No template strategy for outbound
Fix: Submit core templates during build phase, not launch week.

Mistake 3: Unbounded LLM replies on regulated topics
Fix: Scope knowledge base, add guardrails, require human escalation on sensitive intents.

Mistake 4: Ignoring opt-out and data retention
Fix: Document consent, honour stop requests, align storage with DPDPA.

Mistake 5: WhatsApp silo with no CRM write-back
Fix: Every qualified conversation creates or updates CRM within seconds.

Mistake 6: Using personal WhatsApp for business scale
Fix: Migrate to API before volume makes migration painful.