How to Build a WhatsApp Chatbot Without Code: Step-by-Step

This guide is for business owners, marketing managers and support leads who want a working WhatsApp chatbot without writing code. It walks through eight steps from goals to launch, with design limits from Meta, Arabic and English guidance, handover rules and the metrics to track.

What is a WhatsApp chatbot?

A WhatsApp chatbot is an automated assistant that receives customer messages on a WhatsApp Business API number and replies with answers, menus or actions, handing over to a person when needed. The customer sees a normal WhatsApp chat.

There are two kinds:

• Flow (rule-based) chatbot: follows paths you design: welcome menu, buttons, lists, keywords, fixed answers and lookups. Predictable and easy to review.
• AI agent: understands free text, answers from a knowledge base and can use tools. More flexible, and needs stronger guardrails. See WhatsApp AI agents for customer service.

Most teams should start with a flow bot for the top requests and add an AI agent once the basics work.

What do you need before you build?

You need a WhatsApp Business API number, a verified business with a display name, and a few approved templates for messages your business starts. Customer-started chats do not need templates.

1. A WhatsApp Business API number from a provider. The free WhatsApp Business app is a different product; see API vs Business app and how to set up the API.
2. Business verification and display name. Use the readiness checklist to prepare documents.
3. A shared inbox where agents take over. See the Omni Inbox.
4. Content owners. Someone has to approve answers on prices policy, refunds and opening hours.

How do you build a WhatsApp chatbot without code? 8 steps

Follow these steps in order. Each one reduces the risk of the next.

Step 1: Set one clear goal

Pick one job for version one: answer FAQs, qualify leads, book appointments, give order status or collect feedback. Write one measurable target, such as "answer the five most common questions without an agent".

Step 2: Find the top intents

Export or read 100 recent conversations and tag each by topic. The five to ten topics at the top become your intents. Add "talk to a person" as an intent from day one.

Step 3: Write the flow on paper

Draw each path: trigger, bot message, customer choices, outcome. Keep every path under four steps. A customer should reach an answer or a person quickly.

Step 4: Choose the right message format

Use the format that fits the choice, within Meta's limits:

More detail on each format is in interactive messages: buttons and lists. If you have more than ten options, group them into two levels.

Step 5: Write copy in Arabic and English

Write each message natively in both languages and ask for the language in the first reply. Rules that avoid common problems:

• One idea per message; WhatsApp is read on small screens.
• Arabic is right to left; preview variables, numbers and mixed Latin text on a real phone.
• Meta lists Arabic template languages ar, ar_EG and ar_AE; check its list when choosing the template language for Saudi or other markets.
• Keep the tone consistent with your brand, and read Arabic customer messaging best practices.

Step 6: Design the handover

Every path needs a way to reach a person. WhatsApp's business policy expects prompt, clear and direct escalation paths where automation is used, such as in-chat transfer to a human agent. Define triggers:

• The customer asks for a person, or types words such as "agent".
• The bot fails to understand twice in a row.
• The topic is sensitive: a complaint, payment dispute, medical issue.
• A high-value customer or lead is identified.

When the bot hands over, pass the conversation history, tell the customer what happens next and, outside business hours, when a reply will come. Our guide to chatbot to human handoff has the design details.

Step 7: Test with real phones

Test before any customer sees it. A checklist:

1. Every button and list row leads where it should.
2. Typos, voice notes, images and emoji get a polite fallback, not silence.
3. "Agent" works from every screen.
4. Arabic and English versions both display correctly.
5. The bot behaves correctly after the 24-hour window closes: free-form messages are allowed only inside the window; after it closes you can only send approved templates.
6. Lookups return correct data and fail gracefully when your system is down.
7. Test numbers and data are not left in production.

Test access should go to your own verified numbers first, then a small group of staff, then a limited audience.

Step 8: Launch small, measure and expand

Announce the bot on your website, QR codes and ads. Review transcripts weekly for the first month, fix the top three failures and add intents one at a time.