How to Create a WhatsApp Chatbot: Step-by-Step

Building a WhatsApp chatbot is mostly a design job, not a coding one. If you use a platform on the official WhatsApp Business API, you can have a working bot live in a day. Here is the full sequence.

What you need first

• A phone number not currently active on the regular WhatsApp or WhatsApp Business app (or one you can migrate).
• A Facebook Business Manager account.
• A platform that provides WhatsApp Business API access — this article uses BizzBuzz.
• Your FAQs, product or service list, and a rough map of the conversations you get most.

Step 1 — Connect your number to the API

In BizzBuzz, link your Facebook Business Manager and add the number you want the chatbot to run on. The platform provisions it on the WhatsApp Cloud API and configures webhooks and token handling. You then complete Meta's business verification and pick a display name, which Meta reviews.

Step 2 — Map the conversations before you build

Open your WhatsApp inbox and list the last 50 messages by type. You will usually find 5–8 patterns cover most of the volume: order status, opening hours, pricing, booking, returns, “is this in stock”. Those patterns are your first flows. Don't try to automate everything — automate the repetitive majority and route the rest.

Step 3 — Build the welcome flow

The first message sets expectations. A good welcome flow:

1. Greets the customer and says it is an automated assistant.
2. Offers 3–4 buttons for the most common intents.
3. Always includes a “Something else” or “Talk to a person” option.

Keep it to one screen. Long menus kill completion rates.

Step 4 — Build each intent flow

For every button in the welcome menu, build the branch behind it using the visual editor — text, quick replies, list messages, and conditions. Where a flow needs live data (an order, a slot, a price), connect it to your store or calendar so the bot answers with real information rather than a canned line.

Design each flow to end in one of three ways: resolved, a clear next action, or a clean hand-off to a human with context attached.

Step 5 — Add an AI fallback

Rule-based flows can't cover every phrasing. Point an AI agent at your knowledge base and product catalog so that when a message doesn't match a flow, the bot still gives a useful, on-brand answer — and hands off when it is unsure. This is the difference between a chatbot that frustrates people and one that helps them. See chatbot examples for flows worth copying.

Step 6 — Set up the hand-off

Decide when the bot should escalate: on request, after two failed attempts, on negative sentiment, or for high-value contacts. Route escalations into a shared team inbox with the conversation summary, so agents pick up without asking the customer to start over.

Step 7 — Submit templates and test

Any message your business sends first (outside a 24-hour service window) must be an approved template. Submit the templates you need — marketing, utility, authentication — then test every flow end to end on a real phone, including the failure paths.

Step 8 — Launch and watch the first week

Go live on a subset of traffic if you can. Watch three numbers: flow completion rate, hand-off rate, and customer satisfaction after bot-only conversations. Trim menu options nobody taps, and add flows for whatever keeps hitting the hand-off.

Next: see the WhatsApp chatbot product, or read what the WhatsApp Business API is if you are still deciding whether you need it.