How to Build a WhatsApp Chatbot for Your Business in India (2026): The No-Code Guide

Why a WhatsApp chatbot is the highest-leverage automation for most Indian businesses

Your customers already live on WhatsApp — over 500 million Indians use it, and for most small businesses it is where orders, appointments, and support questions actually happen. A chatbot there is not a vanity project. It answers the same ten questions your team answers fifty times a day, captures leads at midnight, and sends payment and appointment reminders without anyone remembering to.

The catch is that most guides are either vendor sales pages or developer tutorials that assume you can run a Flask server. This guide is for the business owner who wants a working chatbot, knows roughly what it costs, and does not want to write code.

The two paths: no-code platform vs Meta’s Cloud API directly

Every WhatsApp chatbot runs on the WhatsApp Business Platform (the API). How you reach it decides how much work and money it takes.

For roughly 9 out of 10 Indian small businesses, the no-code path wins. The Cloud API is genuinely cheaper at scale — no platform fee — but only if someone on your team can own the engineering. A developer tutorial for the code route exists on dev.to if you go that way; the rest of this guide walks the no-code path.

The real cost: two layers, not one

This is where most businesses get surprised. A WhatsApp chatbot has two separate costs, and comparison articles usually show only the first.

Layer 1 — the platform subscription. This pays for the builder, inbox, and automation. Verified against live pricing pages in October 2026:

• AiSensy: chatbot plans at ₹2,500/month for 5 drag-and-drop chatbots (₹2,250/month billed annually); a Smart WhatsApp AI Agent add-on at ₹1,350/month with 1,000 AI messages included.
• Interakt: Growth at $55/month (+taxes) with a basic linear chatbot builder; Advanced at $69/month with branching flows, API calls, and conditions; a WhatsApp AI Agents add-on at $74.99 on those plans. A separate Sales CRM plan runs ₹2,499/month (+taxes) with 5 agents.

All billing on these platforms is prepaid, and plans change often enough that you should re-check the pricing page before budgeting.

Layer 2 — the per-message charges. WhatsApp itself charges for messages, billed through conversation credits on the platform. From 1 October 2026, every business number gets 1,000 free service messages a month; after that, service messages cost ₹0.145 each in India. Marketing templates cost about ₹1.09 per message, utility and authentication templates about ₹0.145 each (AiSensy’s live rates, which include the provider’s markup over Meta’s base rates — see our WhatsApp Business API pricing breakdown for India for the two-layer math in full).

The detail that matters for chatbot economics: bot replies inside the 24-hour customer service window count as service messages — free for the first 1,000 a month. That means a chatbot answering customer questions costs almost nothing in message fees until you cross that threshold. Outbound campaigns (offers, reminders sent as templates that start a conversation) are where the per-message meter runs.

A worked example for a clinic doing 2,000 bot-handled chats a month:

Under ₹3,000 a month to stop missing appointment confirmations. That is the kind of math to do before you build — not after.

Step 0: decide what the bot does before you open any builder

The bots that fail try to do everything. Pick one primary job first: lead capture, appointment booking, order status, or FAQ support. You can add flows later; launch with one.

Then do the unglamorous homework:

1. List your 10 most repeated customer questions. Ask your team or scroll your WhatsApp history. These become the bot’s FAQ answers — the highest-value, lowest-effort automation you own.
2. Define the human handoff. Decide exactly when the bot gives up: a complaint, a pricing negotiation, anything it cannot answer confidently. Every bot needs an escape hatch (“Type AGENT to talk to our team”), or frustrated customers will just stop messaging.
3. Decide your language. Indian customers write in Hinglish — “Order status check karna hai”, “Price kya hai”. Write your bot replies the way your customers type, not in formal English.

Step-by-step: building your first chatbot (no-code)

The screens differ between platforms, but the sequence is the same everywhere. Using AiSensy as the worked example since its pricing is fully transparent on its site:

1. Sign up and connect your number. Register on the platform and connect a dedicated business number to the WhatsApp Business API. The number must not be active on the regular WhatsApp app — once registered for the API, it cannot go back. Platforms procure the API access for you at no extra charge; AiSensy includes a free onboarding call.
2. Complete your business profile. Logo, description, business hours, catalog if you sell products. This is what customers see when they open your chat.
3. Create your first flow. Open the chatbot builder and start with the welcome message — the greeting every new conversation receives. Keep it short and useful: who you are, and three tappable options (pricing, book/buy, talk to a human). A simple 3-option menu handles the majority of inbound queries for most small businesses.
4. Add your FAQ answers. For each of your top-10 questions, create a keyword trigger (“price”, “timing”, “location”, “refund”) mapped to a clear answer. Add buttons and quick replies so customers tap instead of type — fewer dead ends.
5. Set up the human handoff. Configure the fallback: when the bot does not understand a message twice, or the customer types AGENT, route the chat to your team’s shared inbox with a note on what the customer asked.
6. Submit your message templates. Any message that starts a conversation outside the 24-hour window — appointment reminders, order updates, offers — must use a Meta-approved template. Submit them in the platform’s template section; clear, non-spammy templates usually get approved within hours. (More on the rules below.)
7. Test like a customer. Message your own number from a personal phone. Break the bot on purpose: type gibberish, ask off-topic questions, press the wrong buttons. Fix every dead end before launch.
8. Go live, then watch. Launch with your one primary flow. Read the conversation logs for the first week — the questions the bot fumbled are your next flows.

Interakt users follow the same sequence; the main difference is where features sit — linear chatbot flows on Growth, branching flows with API calls and conditions on Advanced, and the AI-agent add-on if you want the bot to answer freely from your website content instead of fixed flows.

The 5 starter flows that earn their keep

Do not invent exotic automations. These five cover what Indian small businesses actually need, roughly in the order to build them:

Once leads flow in, the next leverage is what happens after the chat — piping captured leads into a CRM and follow-up sequences, which is where workflow automation tools pick up the story.

Write like a human, not a menu tree

The difference between a bot customers tolerate and one they use comes down to writing:

• One message, one idea. Long paragraphs get skimmed; customers on WhatsApp read in bursts. Split a 4-line answer into two short messages.
• Buttons beat typing. Every question the bot asks should offer tappable options where possible. Typing is friction; tapping is momentum.
• Hinglish where your customers are Hinglish. If your customers write “Bhaiya price kya hai”, a bot replying in corporate English feels wrong. Match their register — it also sidesteps the uncanny stiffness of translated templates.
• Admit limits fast. “I can help with orders, appointments, and pricing. For anything else, I’ll connect you to our team right away.” A bot that pretends to understand everything and fails is worse than a menu.
• Use AI replies for the long tail, not the core. AI add-ons (AiSensy’s AI Agent, Interakt’s AI agents) can answer from your website or documents when no fixed flow matches. Useful — but keep your money flows (pricing, booking, payments) on deterministic button flows you control. AI is the safety net, not the front door.

For writing better bot replies and prompts, the 6-part prompt framework (role, context, task, format, examples, constraints) works as well for chatbot scripts as it does for ChatGPT.

Template rules you cannot ignore

Meta treats WhatsApp as a messaging service, not an ad network, and it enforces that with template rules:

• Starting a conversation needs an approved template. Inside the 24-hour customer service window you can send anything; outside it, only pre-approved templates in the marketing, utility, or authentication categories.
• Marketing templates cost the most (~₹1.09/message in India) and get the most scrutiny. Utility templates (order updates, appointment reminders, payment receipts) are cheap (~₹0.145) and rarely flagged — which is why smart businesses design their flows to trigger customer replies that open the free service window.
• Opt-in is not optional. Customers must agree to receive messages. Broadcasting offers to a purchased contact list is the fastest route to a restricted number.
• Every promotional flow needs an opt-out. “Reply STOP to unsubscribe” is not politeness; it is how you stay out of the spam reports that trigger restrictions.

The mistakes that get business numbers restricted

1. Importing a cold contact list and broadcasting offers. The single most common cause. Meta’s systems and user reports catch this fast.
2. Skipping template approval and sending free-form promotional messages outside the service window.
3. No human fallback. A bot that traps angry customers with no way out generates the exact spam reports that get numbers limited.
4. One bot for everything on day one. Complex flows break, customers get confused, and you abandon the project. Start with one flow, prove it, expand.
5. Ignoring the 24-hour window. Designing campaigns as if every message were free, then being surprised by the template bill. Map which messages fall inside the window before you launch.

Measure these 5 numbers, ignore the rest

Vanity metrics will tell you the bot is “working”. These will tell you the truth:

1. Containment rate — what share of conversations the bot resolves without a human. Aim for 60–70% on FAQ-heavy flows; below 40% means your flows do not match real questions.
2. Handoff rate and reason — how often the bot gives up, and on what. Your roadmap lives here.
3. Response time to first human reply — the bot’s job is instant acknowledgement; the team’s job is fast follow-up when it hands off.
4. Booking/lead conversion — did the bot actually capture the lead or complete the booking? A bot that chats but never converts is entertainment.
5. Cost per resolved conversation — platform fee plus message charges divided by conversations the bot handled. Compare it honestly against a part-time hire’s cost per chat.

FAQ

Do I need the WhatsApp Business API to run a chatbot? Yes. The free WhatsApp Business app does not support chatbots or automation at scale. Chatbots run on the WhatsApp Business Platform (the API) — either directly through Meta’s Cloud API or via a provider like AiSensy or Interakt that manages the API connection for you.

How much does a WhatsApp chatbot cost in India? Two layers: the platform subscription (around ₹2,500/month for an entry no-code chatbot plan, verified October 2026) plus WhatsApp’s per-message charges for templates sent outside the customer-service window (about ₹1.09 per marketing message, ₹0.145 per utility or service message in India). Bot replies inside a customer-initiated 24-hour window are free for the first 1,000 service messages each month.

Can I build a WhatsApp chatbot without coding? Yes. No-code platforms offer drag-and-drop flow builders, keyword triggers, shared inboxes, and AI-reply add-ons. The direct Cloud API route is cheaper at scale but needs a developer to build and maintain the bot logic, hosting, and webhook handling.

Do chatbot replies cost money? Replies inside the 24-hour customer service window count as service messages: the first 1,000 per month per number are free, then ₹0.145 each in India (rates from October 2026). Messages that start a conversation outside the window are charged at the marketing or utility template rate.

Why did my WhatsApp business number get restricted? The usual causes are sending promotional messages without opt-in, using unapproved templates, or accumulating spam reports from recipients. Keep to approved templates, maintain a clear opt-out path, and warm up new numbers gradually instead of blasting on day one.

Should I use a no-code platform or Meta’s Cloud API directly? Choose a no-code platform if you have no developers and want to be live this week — faster setup, shared inbox, template management included. Choose the direct Cloud API if you have engineering capacity, send very high message volumes (platform fees approach zero), or need custom integrations a platform cannot offer.

Sources

• AiSensy pricing page (aisensy.com/pricing) — chatbot plan pricing, Smart AI Agent pricing, India per-message rates, 1,000 free service messages/month, Oct 2026 pricing update. Verified live 2026-10-08.
• Interakt pricing page (interakt.shop/pricing) — Growth/Advanced plan pricing, chatbot builder tiers, WhatsApp AI Agents add-on pricing, Sales CRM plan. Verified live 2026-10-08.
• Meta WhatsApp Business Platform documentation (developers.facebook.com) — template categories, approval requirements, and the 24-hour customer service window rules.
• KnowledgeSense: WhatsApp Business API Pricing in India (2026) — Meta’s India per-message rates and the two-layer cost model.
• KnowledgeSense: Best AI Voice Agents in India (2026) — when the job needs a voice call instead of a chat.

AI-assistance disclosure: this article was researched and drafted with AI assistance and reviewed against primary vendor pricing pages and Meta’s platform documentation before publication. Pricing figures were verified on 2026-10-08 and change frequently — always confirm on the vendor’s pricing page before budgeting.