I have lost count of the customer portals I have been shown in this region that almost nobody logs into. The client paid for it, the design is fine, and the usage figures are quietly embarrassing. When we go and ask the customers why, the answer is never about the software. They already have WhatsApp open. They were never going to remember a password for something they touch twice a month.
So the sensible move is to meet them where they are. That part is now widely agreed. What is less understood is what building there properly involves.
A number somebody watches is not automation
Most businesses here already use WhatsApp with customers. Usually it is a phone on a desk, or a shared number, and a person reading messages and then typing the same information into the order system afterwards.
That is a second inbox, not a system. It has all the cost of a channel and none of the benefit. Worse, it creates two versions of the truth: what the customer was told in the chat, and what the records say. When those disagree, and they will, nobody can prove which came first.
What earns its cost is when the conversation changes your records without a person in between. A customer asks where an order is and the answer is read out of the system as it stands right now. They confirm a collection date and the date changes in the order. An enquiry comes in and a customer record exists before anyone has looked at it.
We built a sales system in which the software itself asks the first round of qualifying questions inside WhatsApp and Telegram and writes what it learns straight into the customer record. For a tailoring business, order updates go out when the stage of the work changes in the system, not when somebody remembers to send them. In both cases the useful part is the connection back into the records, not the chat.
The rules, which people find out about too late
WhatsApp is not an open channel you can send into. It is run as a business platform with conditions attached, and they are not negotiable:
- Your business account has to be approved and verified, and the phone number you use cannot be one already running on the ordinary WhatsApp app on somebody's handset.
- The customer has to agree to be messaged first. Buying a list and sending to it is the fastest way to get your number shut off.
- Once a customer messages you, you can reply freely for a limited period. Outside that window you may only send wording you submitted and had approved in advance.
- Those approved messages are charged per send, the price varies by country and by what kind of message it is, and the pricing has been changed more than once.
- Approval of the wording takes days, and messages get rejected for reasons that are not always obvious from the outside.
None of that is an argument against doing it. It is an argument for finding it out in week one rather than the week before launch. Business verification alone can hold a project for a fortnight, and it is nearly always started too late because everybody assumed it was a formality.
Where it stops being the right channel
Anything with a long form. Anything where somebody needs to compare options next to each other. Anything where you will later need to prove exactly what the person saw and agreed to, because a chat thread is a poor record for that.
And anything where the answer must be right first time with nobody to catch it. If the software is going to quote a price or confirm a delivery date, decide in advance which of those it may send on its own and which need a person to press send. Get that boundary wrong and you will spend more time apologising than you saved.
Telegram, and why we usually build for both
Telegram is cheaper to build on, the restrictions are lighter, and you are not waiting on anybody to approve your wording. It has a smaller share of customers here but it is well suited to internal use, where your own staff are the ones receiving updates and you control which app they have.
If you are nervous about committing, start there. The work of connecting a chat channel into your records is mostly the same either way, and the second channel is far cheaper than the first.
One thing to look at this week
Find the information your team currently retypes out of a chat into a system, or out of a system into a chat, every single day. That is the piece worth building first, because you already know its volume and you already know what it costs you.
If you want to know what your case would involve, send me a description of that one journey and I will tell you what is straightforward, what runs into the messaging rules, and what is not worth doing at all.
Working through this on a live programme?
A 45-minute call with the engineer who would run the work. We will tell you whether AI is the answer, including when it is not.