How to build a WhatsApp chatbot without code

It's 11:40pm. A customer messages your store asking "where's my order?" Nobody's awake. In the old world that chat sits unread until morning and the customer stews. With a no-code bot, they get their tracking link in four seconds and go to bed happy. That's the whole promise here — and you can build it this afternoon without touching a line of code.

This guide walks through the actual build: the types of bot, what to put in each step, when the bot should shut up and fetch a human, and two working examples you can copy.

Three kinds of no-code bot (you'll use a mix)

Before you build anything, know the three building blocks. Real bots stitch them together.

• Keyword triggers. The customer types a word — "price", "refund", "track" — and the bot fires the matching reply. Fast to set up, but it only works if you guess the words people use.
• Button and list menus. Instead of hoping the customer types the right word, you show tappable options: Track order · Talk to sales · Return an item. This is the workhorse. Buttons remove guesswork, cut typos, and keep the conversation on rails.
• AI-answered. For the long tail of questions a menu can't predict ("does this kurta run small?"), an AI layer reads your FAQs and products and answers in plain language. We cover that in depth in can an AI chatbot handle support 24/7?

Honestly, most of the value is in buttons. They're predictable, they're cheap, and customers actually prefer tapping over typing. Start there; add AI once the rule-based bot is solid.

Step 1 — Write the top 5 questions down first

Don't open the builder yet. Open your inbox instead. Scroll the last two weeks of chats and count what people actually ask. For most Indian stores it's some version of: where's my order, what's the price, do you deliver to my pincode, how do I return this, and is this in stock. Those five questions are your bot. Everything else is noise you'll route to a human.

This one habit is what separates a bot that deflects real work from a bot that annoys people. You're not building a personality. You're building answers to the questions that already eat your team's day.

Step 2 — Build the welcome + main menu

The first message a customer gets sets the tone. Keep it short, name yourself, and immediately offer choices instead of an open-ended "How can I help?" that invites a wall of text.

A good welcome looks like this: a one-line greeting, then a button menu of your top 3–4 intents. Something like:

Hi! You're chatting with Rangoli Apparel 🌸 What can I help with today?
[Track my order] [Sizes & availability] [Returns] [Talk to a human]

Notice the last button. Always give people a visible escape hatch to a person. Bots that trap customers in a loop are the fastest way to earn a block — and blocks hurt your number's quality rating.

Step 3 — Answer each menu branch

Now fill in what each button does. Two of them are self-contained; two need a little logic.

• Track my order → ask for the order number, then reply with the status. If your bot is connected to your store, it pulls the live tracking link automatically. If not, it can still collect the number and hand it to an agent.
• Sizes & availability → this is where an AI answer or a link to your size chart shines, because you can't pre-write every product question.
• Returns → send the return policy and the steps as a short numbered list. Static, reusable, done.
• Talk to a human → straight to hand-off (Step 6).

Keep each answer to a few lines. WhatsApp is a chat, not an email. If an answer needs more than five lines, link out to a page instead of dumping it in the thread.

Step 4 — Office hours and the away message

Set your working hours in the builder. Inside those hours, a "Talk to a human" tap can promise a live reply in a few minutes. Outside them, the bot sends an away message that's honest about timing — "Our team's offline right now, back at 10am. I can still track orders and answer FAQs meanwhile 👇" — and keeps serving the automated branches.

Here's the nice part about after-hours bots: because the customer messaged you first, every reply the bot sends is inside WhatsApp's 24-hour service window, which means those messages are free and unlimited. Your bot working the night shift costs you nothing in message fees.

Step 5 — Routing: send people to the right place

As you grow past one agent, "Talk to a human" shouldn't dump everyone in one pile. Route by intent. A "Returns" escalation goes to support; a "Talk to sales" tap goes to the sales queue; a pricing question about bulk orders goes to whoever handles wholesale. In a shared team inbox, the bot assigns the chat to the right team or agent automatically, so nobody's picking through a shared queue.

Step 6 — The clean hand-off (this is where most bots fail)

A bad hand-off makes the customer repeat everything to the human. A good one carries the context across. Before it hands off, your bot should quietly collect the essentials — order number, the question, maybe a category — and attach them to the chat so the agent opens it already knowing what's going on.

The hand-off itself should feel smooth: the bot says "Connecting you to Priya from our team, one moment," pauses its own replies for that conversation, and pings the assigned agent. When the agent is done, the bot can quietly resume for the next new question. The customer never feels the seam.

When the bot should NOT try to answer

Restraint is a feature. Program your bot to step back and fetch a human whenever:

• It's unsure — low confidence beats a confident wrong answer every time.
• The same question fails twice. Two misses and a human takes over. Don't loop.
• Money is involved — refunds, cancellations, disputed charges, "I was overcharged."
• The tone turns upset or urgent. An angry customer wants a person, now.
• It's a complaint, a legal or safety issue, or anything a wrong answer could make worse.

The test is simple: if a wrong automated answer would cost you money or trust, the bot shouldn't attempt it. It should collect the details and route.

Two real examples

1. The store order-status bot

The most common bot, and the highest-value one. Customer messages after a delivery is late. Bot greets, offers the menu, they tap "Track my order," bot asks for the order ID, looks it up, and replies: "Order #4821 is out for delivery, arriving today by 7pm — here's live tracking 📦." If the order's genuinely stuck, the bot doesn't bluff; it says "Let me get a teammate on this" and hands off with the order number attached. This single flow can quietly clear a big share of your "where is my order" volume, day and night. Pair it with abandoned-cart recovery and the same number does double duty.

2. The lead-qualifying bot

Great for real estate, coaching, B2B, high-ticket services. A click-to-WhatsApp ad drops a lead into the chat. The bot welcomes them and asks three quick button/short-answer questions — what are you looking for, budget range, city — then tags the contact and routes hot leads straight to a salesperson while nurturing the rest. The team stops spending its morning asking "so what's your budget?" fifty times; the bot did the boring part and only the ready-to-buy leads reach a human.

Rule-based vs AI: a quick comparison

You don't pick one. You start rule-based, get the menus and hand-off solid, then bolt on AI for the questions your buttons can't cover.

Test it before you turn it on

Message the bot yourself as if you were a confused customer. Type nonsense. Tap the wrong button. Ask something off-script. Every path should either answer or reach a human — no dead ends, no loops. Then watch the first week's real chats and fix wherever people got stuck. A bot is never "done"; it's tuned.

The bottom line

A no-code WhatsApp chatbot isn't about replacing your team — it's about deleting the repetitive 60–70% so your people handle the chats that actually need a human. Map your top five questions, build a button-first flow, respect your office hours, and obsess over a clean hand-off. Do that and your worst-case customer experience — a message at midnight — becomes one of your best.