Nigerian businesses run on WhatsApp — vendors, salons, food sellers, boutiques. It's where the customer already is, so it makes sense as a sales channel. The problem is scale: one person can only type so fast, and by the time a customer's third question gets answered, they may have already ordered from someone who replied faster.
An AI ordering bot doesn't replace that WhatsApp presence — it sits behind it, replying instantly, taking the order details, and only handing off to a human when something genuinely needs one.
What the bot actually does
- Answers instantly. No "will reply in a few hours" — the customer gets a response the moment they message.
- Understands intent, not just keywords. A customer typing "how much for the medium size" gets a real answer, not a canned menu dump.
- Remembers the conversation. If a customer messages, leaves, and comes back later, the bot picks the thread back up instead of starting over.
- Takes the order end-to-end. Product, quantity, delivery details — captured and logged without a human typing any of it.
- Runs while you sleep. Orders placed at midnight are captured just as cleanly as ones placed at noon.
What it's built on
A production WhatsApp AI bot isn't a chatbot toy — it's a small piece of real infrastructure. In practice that means a backend (Node.js/Express is a common, reliable choice), a database to track vendors, products, and orders (Supabase/PostgreSQL works well here), a WhatsApp connection layer (libraries like Baileys), and an AI layer for understanding what the customer actually wants (an LLM API such as Claude). Hosting is typically split — Netlify for anything static, Render or a similar host for the always-on backend process.
For a multi-vendor setup, you also need country/region selection, QR-code onboarding for new vendors, and a dashboard where a vendor can log in and see their own orders — otherwise it's just one bot for one business, not a platform.
Where it actually pays off
Food and beverage vendors
High message volume, repetitive questions ("are you open," "what's on the menu today," "how much is delivery") — exactly the kind of traffic a bot handles best, freeing the owner to focus on actually fulfilling orders.
Service-based businesses
Salons, cleaners, tutors — anyone taking bookings can use the same pattern to confirm availability and capture booking details without back-and-forth typing.
Small retail / boutiques
Product questions, sizing, availability — a bot can answer the repetitive 80% and flag the unusual 20% for a human.
What it won't do
It's not a replacement for a human when something goes wrong — a wrong order, a complaint, a genuinely unusual request. A well-built bot recognizes when it's out of its depth and hands off to a real person instead of guessing. That handoff logic matters as much as the automation itself.
Getting started
You don't need to automate everything on day one. A good starting point is picking the single most repetitive conversation in your WhatsApp — "what do you sell / how much / how do I pay" — and automating just that first. From there it's straightforward to expand into full order-taking and vendor dashboards.
Curious what this would look like for your business?
I built and run a live WhatsApp AI ordering system — happy to walk through what it'd take for yours.
Let's Talk