WhatsApp AI Chatbot Cost in India: Build Price and Running Costs

This guide is for business owners, sales heads and operations managers in India who get a steady stream of WhatsApp messages and want to answer them faster without hiring another two people. By the end you will know what a WhatsApp chatbot can actually do, the difference between a rule-based bot and an AI bot, how the WhatsApp Business API works, what the build and running costs look like in rupees, and what the compliance rules are so your number never gets restricted.

What can a WhatsApp chatbot do for a business

A chatbot on WhatsApp is not a novelty. For most Indian businesses it is the fastest way to reduce repetitive typing by your sales and support staff. The jobs it handles well are the ones that repeat many times a day with small variations.

• Answer FAQs: timings, address, price list, delivery areas, return policy, warranty terms. A pump manufacturer in Rajkot gets the same twelve questions from dealers every week; a bot answers them in seconds at 11 PM.
• Order and dispatch status: the customer types an order number or mobile number and the bot reads the status from your billing or ERP software and replies with the courier link.
• Lead capture and qualification: the bot asks name, city, requirement and budget, then creates a lead in your CRM and alerts the right salesperson.
• Bookings and appointments: clinics, salons, coaching institutes and service centres let customers pick a slot, and the bot sends a reminder the day before.
• Payment links: the bot sends a UPI or payment gateway link for advance, fee or invoice payment and confirms once the payment succeeds.
• Catalogue and quotation: a product list with images and prices, plus a quotation PDF generated from what the customer selected.
• Human handover: the bot passes the chat to a staff member with the full history when it cannot help, or when the customer asks for a person.

The common thread is that the bot is useful only when it is connected to your real data. A bot that says "our team will get back to you" for every question is a form, not a chatbot.

Rule-based chatbot vs AI chatbot: which one do you need

There are two ways to build the brain of the bot, and the price difference between them is large, so it is worth understanding both before you ask for quotes.

Rule-based (menu and keyword) bots

The bot shows numbered options or buttons, the customer taps one, and the bot follows a fixed flow. It is predictable, cheap and easy to test. It struggles when a customer types a full sentence like "my order from last Tuesday has not reached Morbi yet", because it only recognises keywords and menu choices.

AI chatbots using a language model

An AI bot reads the message, works out the intent, pulls the relevant answer from your documents or database, and writes a reply in natural language, including Hinglish or regional-language text if you allow it. It handles messy questions well. The trade-offs are higher build cost, a small per-message AI cost, and the need for guardrails so it does not promise a discount you never offered.

In practice the best answer for most businesses is a hybrid: buttons and menus for the top ten tasks, and an AI fallback for everything else, with a human handover rule. That keeps the cost down and the experience natural.

How the WhatsApp Business API actually works

The green-tick, multi-agent, automated WhatsApp you see from airlines and banks runs on the WhatsApp Business Platform, usually called the WhatsApp Business API. It is different from the free WhatsApp Business app on your phone. The app is for one person replying manually; the API is for software replying on behalf of your business number.

1. Business verification: your company is verified through Meta's business manager using registration documents. This is what lets you send messages at scale.
2. A dedicated number: the API number cannot be logged in on a phone at the same time. Most businesses take a new number or migrate an existing one carefully.
3. Access route: you connect either directly to Meta's hosted cloud API or through an authorised solution provider. Either way, your chatbot software talks to the API and the API delivers messages.
4. The 24-hour window: when a customer messages you, you can reply freely for 24 hours from their last message. Outside that window, you can only send pre-approved template messages.
5. Template messages: fixed-format messages such as "Your order {1} has been dispatched via {2}" that Meta approves in advance. They are grouped into categories such as utility, marketing and authentication, and each category is billed differently.
6. Quality rating: if too many people block or report your number, your rating drops and your daily sending limit falls. Opt-in and relevance matter.

Your chatbot lives on your own server or cloud account and connects to the API on one side and to your CRM, ERP, billing or booking system on the other. That integration layer is where most of the build effort goes.

How much does a WhatsApp chatbot cost in India

These are one-time build bands we see for custom bots in Gujarat and across India. They assume the official API, a proper admin panel to edit answers and see conversations, and testing with your team before launch.

Scope	What is included	Build cost (one-time)	Typical timeline
Starter rule-based bot	Menu flows, 20 to 40 FAQs, lead capture to email or sheet, human handover, basic dashboard	₹40,000 to ₹1 lakh	2 to 3 weeks
Business bot with integrations	Everything above plus order status from ERP or billing, appointment booking, payment links, template campaigns, CRM sync	₹1 lakh to ₹2.5 lakh	4 to 6 weeks
AI chatbot	Language-model answers trained on your documents and product data, multilingual replies, intent detection, guardrails, analytics, hybrid menus	₹1.5 lakh to ₹4 lakh	5 to 8 weeks
Enterprise or multi-branch	Multi-agent inbox with roles, branch routing, SLA reports, catalogue and cart, deep ERP and payment integration, audit logs	₹4 lakh to ₹10 lakh and above	8 to 14 weeks

Two things move the price more than anything else: the number of systems the bot has to talk to, and how much of the conversation is free-text AI versus buttons. A bot that reads from one billing database is cheap. A bot that reads from three legacy systems with no APIs is not.

Running costs: per-conversation charges, AI usage and hosting

The build is only part of the bill. Plan for these monthly costs before you commit.

• Meta message charges: Meta bills business-initiated template messages per message or conversation, with different rates for marketing, utility and authentication categories. Customer-initiated service conversations are treated far more generously. Rates change, so check the current rate card for India on Meta's official documentation. As a planning figure, a business sending 5,000 utility notifications a month spends a few thousand rupees; marketing broadcasts cost considerably more per message.
• Solution provider or platform fee: if you go through an authorised provider rather than the direct cloud API, there is usually a monthly platform fee or a small markup per message.
• AI usage: language-model calls are billed per token by the model provider. For a support bot handling 10,000 messages a month with short answers this is typically ₹1,500 to ₹6,000, more if you use a premium model or long documents.
• Hosting: a small cloud server or serverless setup for the bot and its database, ₹1,000 to ₹5,000 per month.
• Maintenance: template updates, new flows, model tuning. Budget 10-15% of the build cost per year or a small monthly retainer.

Put together, a mid-size business should plan ₹3,000 to ₹20,000 per month in running costs. That is usually less than the cost of one part-time support person, and the bot works on Sundays.

What JK Tech Hub has seen on real WhatsApp bot projects

A few patterns from bots we have built for businesses in Rajkot, Ahmedabad and Surat.

A kitchenware manufacturer in the Rajkot industrial belt had two staff members spending most of the morning replying to dealers asking "has my order been dispatched". We connected a rule-based bot to their existing dispatch module. Dealers now type their order number and get the transporter name and LR number. The two staff members moved to following up on pending payments, which had a far bigger effect on cash flow than anyone expected.

A coaching institute in Rajkot used a booking bot for demo lectures. The surprise was that most enquiries came between 9 PM and midnight, after parents finished work. Before the bot those messages waited until 10 AM, and by then a share of parents had already spoken to another institute.

A ceramic exporter in Morbi wanted an AI bot to answer product questions in English, Hindi and Gujarati. The AI part worked well within a week. What took time was cleaning their product catalogue, because the bot can only be as accurate as the data it reads. Budget time for data cleanup on any AI bot project.

The biggest failure we have seen was not technical. A retailer sent marketing templates to a purchased list without opt-in, the number was reported by many recipients, and their sending limit was cut sharply. Rebuilding the quality rating took weeks.

Compliance, opt-in and keeping your number safe

WhatsApp is strict about consent, and India's Digital Personal Data Protection Act adds its own duties around collecting and storing personal data. Build these into the bot from day one.

1. Collect opt-in: customers must agree to receive messages from your business, through a website checkbox, a missed-call flow, a QR code, or by messaging you first. Record when and how consent was given.
2. Offer an easy opt-out: a "STOP" keyword that actually stops marketing messages. Honour it instantly.
3. Use the right template category: do not disguise a promotion as a utility message. Meta reclassifies and may reject templates that do this.
4. Respect the 24-hour window: the bot must switch to templates automatically when the window closes, or the messages silently fail.
5. Protect personal data: store only what you need, encrypt the database, restrict staff access to conversations, and have a retention rule. Check the current DPDP Act rules and obligations on the official government portal, because implementation timelines have been announced in phases.
6. Keep a human path: customers should always be able to reach a person. This is good for the quality rating and for your reputation.

How to decide and what to ask a chatbot developer

Start with a list of the ten most repeated conversations your staff handle. If eight of them are answerable from a menu, a rule-based bot at ₹40,000 to ₹1 lakh is the right first step. If most enquiries are open-ended product questions, budget for AI from the start, but still keep buttons for the common tasks.

When you speak to a chatbot development company, ask these questions: which systems will the bot connect to and how; who owns the WhatsApp number and the Meta business account (it should be you); can your staff edit answers without calling the developer; what happens when the bot does not know; and what the monthly running cost will be at your expected volume. A good vendor will give you a written scope with flows listed one by one.

Finally, measure. Track first-response time, share of conversations closed without a human, leads captured, and payments collected through the bot. A well-scoped WhatsApp chatbot usually pays back its build cost within three to six months on time saved and leads that would otherwise have gone cold overnight.

Related reading and next step

• AI chatbot development for WhatsApp, web and apps
• AI agent development for automating business workflows
• Custom CRM software to receive the leads your bot captures
• Payment gateway and UPI integration for payment links
• A practical guide to integrating language models into business software

Send us your ten most common customer questions and we will reply with a fixed-price quote and a suggested flow within two working days through the contact page or on WhatsApp at +91 7265004040.