What a WhatsApp chatbot in South Africa should qualify before you reply

A WhatsApp chatbot for a South African business earns its cost by qualifying every lead before anyone on your team replies. It asks the questions that decide whether a lead deserves your time and passes the conversation to a person once a real decision is needed. The answers belong in your own system, not only in a vendor's inbox.

What should a WhatsApp chatbot ask a lead before you reply?

A WhatsApp chatbot should ask only the questions whose answers change what you do next with a lead. Every extra question is one more message where a person can stop replying.

The right questions depend on the business. As general practice, a rental business needs area, budget and move-in date, while a services firm needs scope, timing and who signs off. Look at the last twenty leads that wasted your time and ask which single fact would have told you sooner.

DataReportal's Digital 2026: South Africa report counts 51.7 million internet users in South Africa in October 2025, an internet penetration of 79.6 percent. A practical lead qualification flow on WhatsApp looks like this:

1. Write down the facts that decide whether a lead goes further.
2. Turn each one into a question with reply buttons or a short list, so the answer is clean data.
3. Set the rule for which answers get a reply today, a self-booked slot or a polite no.
4. Decide where the answers are stored before you build anything.
5. Decide the exact point where a human takes over.

Where should WhatsApp chatbot answers land: your own backend or a vendor inbox?

WhatsApp chatbot answers should land as structured data in a system you own, because a lead that only lives in a vendor's chat inbox cannot be sorted or reported on, and it stays behind when you change tools. Most guides compare bot builders and never ask what happens to the answers afterwards.

A hosted platform with a shared inbox is fine when one person handles a few leads a day. It stops being fine when the answers need to drive a booking or a report on which ads bring good leads. Then you need database fields, not chat lines.

We built this for a property business run by two founders, working as their embedded CTO over five and a half months. Their WhatsApp chatbot flow handles lead intake and qualification before either founder gets involved. The qualification answers arrive as payloads at an event-driven webhook layer we built in Express, which handles them in real time, and the founders' operations dashboard stays in real-time sync with the WhatsApp side. Automating that top of the funnel is what let the founders step away from manual triage.

Cost is the other reason to keep the flow tight. Since 1 October 2026, Meta's WhatsApp Business Platform pricing page says each business phone number gets 1,000 free service messages per month, and Meta charges from the 1,001st. Non-template replies a third-party chatbot sends inside the 24-hour customer service window count toward that, so a ten-question flow uses far more of the allowance than a four-question one. Our breakdown of WhatsApp Business API pricing and session windows explains how that window works.

Where should a WhatsApp chatbot stop and a person take over?

A WhatsApp chatbot should hand over the moment the next message needs a person's judgement, which means a lead that qualifies or a question the flow was not built for.

Show a way to reach a person at every step, say when a human will reply, and never let the bot pretend to be one. Ask your developer or vendor four things. Where are the answers stored, and can I export them? What happens when someone types something the flow does not expect? How many service messages does one full qualification use? Who owns the WhatsApp number if we stop working together? If WhatsApp might become your whole customer front end, read whether your business needs an app or WhatsApp. When the answers must feed your own dashboard or booking system, that is a custom software build, not a plugin.

If you get more WhatsApp enquiries than you can answer properly each day, a qualification bot is worth building, and the real decision is where its answers go. If you get a handful a week, the WhatsApp Business app and your own replies will do.

Questions about WhatsApp chatbots

How do you qualify leads on WhatsApp automatically?

You qualify leads on WhatsApp automatically with a chatbot flow on the WhatsApp Business Platform that asks a few fixed questions, mostly as reply buttons, and checks the answers against rules you set. Qualified leads go to a person or a booking step. In the property system we built, those answers reached the founders' backend in real time.

How much does a WhatsApp chatbot cost in South Africa?

A WhatsApp chatbot in South Africa has two costs: the software, either a platform subscription or a custom build, and Meta's message fees. A subscription is cheaper to start. A custom flow costs more upfront but stores answers in your own system. Since October 2026, Meta charges for service messages beyond 1,000 per number each month.

Do WhatsApp chatbot replies cost money since October 2026?

Yes, beyond a free allowance. Meta's pricing page says each business phone number gets 1,000 free service messages per month, which reset monthly and do not roll over, and Meta charges from the 1,001st. Non-template replies a third-party chatbot sends inside the 24-hour customer service window count as service messages, so a shorter qualification flow uses less.

When should a WhatsApp chatbot hand a lead to a human?

A WhatsApp chatbot should hand a lead to a human once the lead qualifies or asks something the flow was not built for. Offer a visible way to reach a person at every step and tell the lead when to expect a reply.

Does POPIA apply to a WhatsApp chatbot that collects lead details?

Yes. A name, phone number and answers about someone's needs are personal information under POPIA, so the business collecting them needs a lawful reason, must tell the lead what it collects and why, and must keep it secure. If a chatbot vendor stores data outside South Africa, check how that transfer is handled, and take legal advice.

What is the difference between a rule-based WhatsApp chatbot and an AI agent?

A rule-based WhatsApp chatbot follows a fixed flow of questions and buttons, so its answers are predictable and easy to store. An AI agent writes its own replies and handles open questions, but it is harder to control. In the property system we built, qualification ran as a chatbot flow, and AI was used for property matching.