The WhatsApp automation that gets your number banned

Both get sold as "WhatsApp automation" and they are not the same category of thing.

The first is the official WhatsApp Cloud API from Meta. You verify your business, you get a verified business number, and you send and receive messages through an interface Meta built for exactly this. It is unglamorous and it stays switched on.

The second is a tool that automates WhatsApp Web — software that drives the browser interface the way a very fast person would. It is cheaper, it sets up in an afternoon, and it does not require verifying anything. It is also against WhatsApp's terms, and the enforcement is not a warning.

Numbers running unofficial automation get banned. Not always immediately — that is the trap. The pattern I see is that it works fine for months, the business quietly builds its whole enquiry flow around it, and the ban lands once there is real volume going through it.

And it is the number that gets banned. Not the software. The number on your shopfront, your van, your Instagram bio, your printed cards — the one your existing customers have saved. You do not lose an automation tool that day, you lose the address your business is reachable at.

That is why the price difference is not really a price difference. You are being quoted for two different risk profiles and only one of them is on the invoice.

This is the part people find genuinely surprising, so it is worth being specific.

Meta bills per conversation, not per message. A conversation is a 24-hour window, so a customer who sends you fifteen messages in an evening is one conversation, not fifteen. The rate depends on category and country: service conversations, the ones your customer starts, are the cheapest. Marketing ones you start cost more.

For most small businesses this lands in the low thousands of rupees a month. That bill comes to you directly from Meta, and it is separate from whatever you pay whoever builds the thing.

The setup work — business verification, the Cloud API number, message templates submitted for approval — is mostly paperwork, and it is the part worth paying someone to do because the template approval process is fussy in ways that are not documented anywhere useful.

The instinct is to build a bot that can handle everything. Do not. Go and read your last two weeks of real WhatsApp chats instead, and count.

It will be the same six questions, over and over: what does it cost, do you have it in stock, where are you, what are your timings, where is my order, do you deliver to my area. Those six are most of your message volume and none of them need a person.

Automate exactly those, and what reaches your team is the conversation actually worth their time. That is the whole return, and you get it in the first fortnight — long before anything clever.

A bot that cannot answer has two options and only one of them is survivable.

It can guess, which means one day it quotes a price you do not honour or promises a delivery date you cannot meet. Now you either eat the cost or spend the conversation explaining why your own system was wrong. That bot has cost you more than having no bot at all.

Or it can stop, say a person will pick this up, and hand over with the whole conversation attached so the customer does not have to repeat themselves. That is the bit worth building carefully, and it is the bit most cheap builds skip, because it is invisible in a demo.

One more thing, since it comes up every time: yes, customers should be able to tell it is a bot, and no, they do not mind. Pretending otherwise backfires the moment it slips — and nobody who got a real answer at 2am has ever complained that it was automated.