You don't need a developer to put a bot on WhatsApp anymore. With an AI-powered no-code builder, you can go from idea to a live WhatsApp bot in under an hour. In this tutorial, we'll build a customer support bot for a small shop complete with button menus, an AI agent for free-form questions, and an M-PESA payment step.
What You'll Need
- A free UFlow account
- A WhatsApp number for your business (any number that can receive WhatsApp)
- 10–60 minutes, depending on how fancy you get
Step 1: Generate Your Bot with AI
After signing in, click "Create New App" and choose WhatsApp Bot. Instead of starting from a blank canvas, open the AI assistant and describe what you want:
"A customer support bot for my electronics shop. Greet the customer, show a menu with Products, Order Status, and Talk to Support. Products should show a list of categories. Order status should ask for an order number."
In seconds, the AI generates the complete flow greeting message, button menu, list menus, input steps, and all the connections between them.
Step 2: Make It Feel Like WhatsApp
WhatsApp gives you much richer building blocks than plain text. In the drag-and-drop editor, you can use:
- Button menus — up to 3 tappable reply buttons per message
- List menus — scrollable lists, ideal for product catalogs
- Media messages — product photos, PDF price lists, location pins
- Typing indicators — short "typing..." delays that make the bot feel human
Pro Tip: Buttons beat typed numbers
On USSD, customers type "1" or "2". On WhatsApp, use reply buttons and lists instead they're faster, eliminate typos, and look professional.
Step 3: Add an AI Agent for Free-Form Questions
Menus handle predictable journeys, but customers love asking questions in their own words. Drag an AI Agent node into your flow and give it context about your business your products, prices, policies, and tone.
Now when someone asks "do you have the Samsung A15 in stock and how much?", the AI answers naturally. You can configure it to hand off to a human agent whenever the conversation goes beyond its knowledge.
Step 4: Collect Payments with M-PESA
To take payment inside the chat, add a Payment node. When the customer confirms their order, UFlow triggers an STK push on their phone. Your flow then branches on the result:
- Success → send a receipt and confirmation message
- Failed → offer to retry or talk to support
Step 5: Test in the Simulator
Before linking a real number, open the WhatsApp simulator built into UFlow. It renders your bot exactly as customers will see it buttons, lists, media and all so you can click through every path and polish the copy.
Step 6: Link Your WhatsApp Number and Go Live
From your app's dashboard, click "Connect WhatsApp". You'll see a QR code scan it with the WhatsApp app on the phone that holds your business number (or use a pairing code). That's it. Your bot is now answering real customers.
No WhatsApp Business API application. No waiting for Meta verification. Live the same afternoon.
Bonus: Reach Feature Phones Too
Not all your customers are on WhatsApp. With UFlow you can clone the same logic into a USSD app that works on every phone in Kenya sharing the same datasets and M-PESA configuration. One platform, every customer.
Start Building Your WhatsApp Bot
Free to build, test, and link. Describe your bot to the AI and watch it come to life.
Try the WhatsApp Bot Builder