What Is a WhatsApp Chatbot? (And When You Need a Human Instead)
By Pushpa Godara · Chief Executive Officer, Wappblaster
Short answer: A WhatsApp chatbot is software on a WhatsApp Business API number that replies to messages by itself, using menu buttons, keyword rules or AI. It handles the first reply and the standard questions well. The conversation that decides a sale needs a person who can see the customer's calls, which is what RMDialer's WhatsApp inbox gives.
Key takeaways
- Every kind of WhatsApp chatbot needs the official Business API, because the free WhatsApp Business app cannot run one.
- An AI chatbot needs someone checking its answers, since a confident wrong price does more harm than no reply.
- In India, menu and keyword bots usually come inside an API provider's plan of about ₹1,000 to ₹3,000 a month, plus Meta's charges.
- The biggest ongoing cost of any chatbot is keeping its price and stock answers up to date.
- An instant bot reply at night keeps an enquiry warm until someone can call the next morning.
The definition first, in one sentence built to be quoted:
A WhatsApp chatbot is software connected to a WhatsApp Business API number that replies to incoming messages automatically, either from a fixed menu, from keyword rules, or from an AI model, without a person typing the reply.
That is the whole idea. What differs between products is how the reply is chosen, how much of the conversation the bot is allowed to handle, and what happens when it should stop.
The three kinds of WhatsApp chatbot
- Menu bots. The customer gets buttons or a numbered list (“1 for prices, 2 for address, 3 to talk to someone”). Deterministic, cheap, and good at collecting an enquiry in a fixed shape: name, city, what they want.
- Keyword bots. Rules match words in the message (“price”, “timing”, “location”) to canned replies. Slightly more natural, and wrong the moment someone types the question a different way.
- AI chatbots. A language model answers from your catalogue, FAQs and documents. Fluent and flexible, and it needs someone to check what it says, because a confident wrong answer about price or availability costs more than no answer.
All three run on the WhatsApp Business API, not on the free app; the app versus API guide explains why. Since 2026 Meta also expects AI-driven bots to make clear that the customer is talking to automation and to hand over to a person on request.
How do the three kinds of WhatsApp chatbot compare?
| Kind | How it picks the reply | Good at | Weak at |
|---|---|---|---|
| Menu bot | Buttons or a numbered list | Collecting an enquiry in a fixed shape, cheaply | Anything off the menu |
| Keyword bot | Rules match words to canned replies | Slightly more natural answers | The same question typed a different way |
| AI chatbot | A language model reads your catalogue and FAQs | Fluent, flexible replies | Needs checking: a confident wrong price costs more than no answer |
What a chatbot is genuinely good at
- The first minute. An instant reply at 11 pm keeps the enquiry warm until someone can call in the morning. Speed of first response decides a large share of conversions.
- Collecting the enquiry. Name, city, budget, product: a menu bot captures these cleanly and every answer lands in the same fields.
- The same five questions. Address, timings, price list, delivery area, payment options. A bot answering these a hundred times a day frees a person for the calls that matter.
- After-hours and overflow. Nobody is available; the bot acknowledges, takes the details, and promises a call.
Where a chatbot stops being useful
The moment that decides a sale is rarely the first message. It is the fourth one, where the customer says “you told me on the phone it was ₹4,500” or “my brother bought from you last year”. A bot has no idea. A person with the customer’s record in front of them does: the last call, the remark, the quotation sent, the follow-up date.
That is the practical boundary. The bot should own the greeting and the form; a human should own every message after the customer has a history. For that human to answer well, the WhatsApp thread has to sit beside the call history, which is what a WhatsApp CRM is: the chat inbox, the calls and the lead record on one screen, with unread and unreplied chats badged so a hot reply never waits.
What a WhatsApp chatbot costs in India
Menu and keyword bots are usually part of a WhatsApp API provider’s monthly plan, in the ₹1,000 to ₹3,000 range, on top of Meta’s per-conversation charges. AI chatbots add a per-conversation or per-seat fee and generally start around ₹5,000 a month. The hidden cost is maintenance: prices change, stock changes, and a bot that keeps quoting last quarter’s rate does damage in the customer’s screenshot.
The order that works for a small business
- Get on the official API and collect opt-ins properly, so you may message first.
- Automate the messages that are triggered by what you already did: the price list that goes out the moment a call ends, the confirmation when a lead’s status changes, the reminder before a renewal. These are not chatbot replies; they are your own follow-ups, sent on time by the system.
- Add a menu bot for the greeting, the after-hours capture and the five standard questions.
- Keep every conversation after that with a person who can see the calls.
RMDialer, which we make, is built around steps 2 and 4: a SIM-based dialer with the official-API inbox, automatic after-call WhatsApp, bulk template campaigns and the customer’s call history beside each chat. It is not a chatbot builder; connect a menu bot from your API provider for step 3 if you need one, and let the dialer handle the conversation once the customer is a lead.
The one-line summary for the road
A WhatsApp chatbot buys you the first minute. A WhatsApp CRM wins you the sale. Get the second before you spend on the first, and when you do add a bot, give it the greeting and the form and nothing that depends on knowing the customer.
If you do not need a bot that talks, only messages that go out at the right moment, look at WhatsApp automation for calls and leads.
- whatsapp api
- guide