Automated WhatsApp Message: How to Set It Up Step by Step (2026)

If you run a business and use WhatsApp to sell or provide support, you already know what happens: messages come in at all hours, many with the same question, and if you don't reply quickly the customer goes elsewhere. Setting up an automatic message seems like the obvious solution, but there's a huge difference between a fixed greeting text and a system that actually answers what people ask. In this guide you'll see both: how to set up the basic automations in WhatsApp Business, how to make the leap to AI-powered replies that match the way you speak, and what to look at before paying for a tool.

What an automatic message on WhatsApp is (and what it isn't)

An automatic message is any reply that fires without you typing at that moment. There are several types worth distinguishing because they're set up differently and serve different purposes:

• Welcome message: sent the first time a contact writes to you or after a period without conversation. It's useful to let them know you're there, give your hours, or ask the person to clarify what they need.
• Away message: activated when you're outside your hours. It's useful so you don't leave the customer hanging and to tell them when you'll reply.
• Quick replies: shortcuts you trigger manually by typing a slash and a word (for example /price). They're not truly automatic, but they save time and are worth having loaded.
• Keyword-based automatic replies: when the customer's message contains a certain word, a reply fires. This already requires an external tool or the WhatsApp API, because the app alone doesn't do it well.
• Time-based auto-replies: in addition to the classic away message, some tools let you define finer rules by day or time slot. It's the same concept, with more control.

Step by step: setting up the welcome message in WhatsApp Business

1. 1Open WhatsApp Business on your phone and go to Settings.
2. 2Find the Business tools section and then Welcome message.
3. 3Turn on the option with the switch at the top.
4. 4Write the text. Keep it short and concrete: who you are, what you offer, what you need the person to do. No long paragraphs.
5. 5Choose whether you want it sent every time someone writes or only the first time. If you send it every time, it becomes annoying; most businesses choose the first time.
6. 6Save and test from another number to see how it arrives.

The away message is set up the same way, in the same section. There it's worth having it sent every time you're outside your hours, because the customer needs to know you're not ignoring them. Include the time you'll be available again and, if you can, a link or clear instruction so they don't keep waiting. A short example: "Thanks for writing. We're outside our hours: we'll reply tomorrow from 9:00. If you want to get ahead, tell us what you need and we'll look at it as soon as we open."

The ceiling of basic automation (and why many businesses stay there)

The greeting and the absence solve the first-impression problem, but not the underlying one: the customer asks something specific (price, stock, shipping, appointment) and still gets no answer. That is where basic automation falls short. You can load quick replies and answer by hand, but if many messages come in per day, the bottleneck stays the same: someone has to read, understand and write.

The alternative that is being used more and more is connecting WhatsApp to a system with artificial intelligence that suggests or generates the response. The key is in how it does it. A traditional flow-based chatbot works with a tree of options that someone has to program and maintain: if the customer writes something that is not in the tree, the bot gets stuck or answers anything. A system that learns from the business's real responses works differently: it looks at how you answer, understands the context and proposes a similar response. The more you use it, the better it responds.

There is a third path, that of numbered-menu bots ("press 1 for sales, 2 for support"). They work for routing, but not for selling or for answering open questions. If your business depends on customers asking different things all the time, that format falls short quickly.

How AI that suggests responses on WhatsApp works

The model that Wando uses, for example, is suggestion-based with human approval: the AI drafts a likely response and the person at the business reviews it and sends it with one click. It does not respond on its own without supervision by default. This is important because in customer service a mistake can be costly: promising a price that is not right, confirming an appointment that does not exist, giving outdated information. With human-in-the-loop, the risk goes down and the team gains speed without losing control.

The other leg is learning. The system feeds on the real responses that the team has been giving, so the suggestions increasingly resemble the business's tone and information. There is no need to build flows or maintain decision trees. That is what makes it different from a classic chatbot. Wando uses Meta's official WhatsApp API (WhatsApp Cloud API), so the channel is the same one you already use, without tricks.

WhatsApp Business API, BSP and the regular app: what each one is

If you're looking for how to automate WhatsApp messages, at some point you'll come across these three acronyms. It's worth understanding them before choosing a tool, because they define what you can and can't do.

• WhatsApp Business (the app): this is the free application you install on your phone. It works for small businesses and allows a welcome message, away message, quick replies and a catalog. It doesn't allow advanced automations or real multi-agent support.
• WhatsApp Business API (Meta's Cloud API): this is the official way to connect WhatsApp to an external system. It allows several people to handle the same line, Meta-approved templates, campaigns and automations. It requires a provider to implement it.
• BSP (Business Solution Provider): this is the provider authorized by Meta to give access to that API. When you read "whatsapp bsp" you're looking for exactly this: who connects you to the official API without you having to do the technical integration yourself.

Common mistakes when setting up WhatsApp automations

• Extremely long messages: the customer doesn't read paragraphs. Two or three lines maximum.
• A greeting that says nothing: "Hi! How can we help you?" adds nothing. Better to say what you do and what you need from the customer.
• Not stating your hours: if you're closed, say so. Otherwise, the customer thinks you're ignoring them.
• Automating everything: there are conversations that need a person. If the bot tries to solve everything, the customer gets frustrated.
• Not measuring: if you don't know how many messages come in outside business hours or how long you take to respond, you can't improve.
• Using the unofficial API: besides the risk of having your account blocked, you don't have the features of the official API or Meta's backing.
• Leaving the same welcome message for years: what worked when you had ten inquiries a week doesn't work when you have two hundred.

How to measure whether your automations are working

There's no single metric, but there are three that help you know if you're on the right track. The first is first response time: if you used to take hours and now you take minutes, the automation is doing its job. The second is how many messages are left unanswered at the end of the day. The third is how many conversations end in a sale or a confirmed appointment. If you have a tool with statistics, these metrics come out on their own; if not, you can keep a simple manual record for a week to have a baseline.

A fourth indicator, more qualitative, is how many times someone on the team had to completely rewrite the AI's suggestion. If they rewrite everything, the tool still hasn't learned the way you speak. If they only correct a detail, it's working.

How to sell products on WhatsApp without losing inquiries

Selling on WhatsApp isn't just about replying quickly: it's about organizing the process. Businesses that sell well through this channel usually have three things figured out. First, an initial response that guides the customer (what you sell, how to buy, what you need from them). Second, a place where contacts are stored with labels, so you don't lose track of someone who asked but didn't close. Third, a system to follow up with those who were left on read, without sounding pushy.

• Set an initial message that already filters: "Tell us what product you're looking for and what area you're in, and we'll send you the price and shipping."
• Save contacts with labels (interested, purchased, waiting for stock) so you can pick up the conversation later.
• Use templates for messages you initiate (follow-up, shipping confirmation), because those require Meta approval.
• Separate customer service from sales: answering a question isn't the same as closing a deal. If multiple people are involved, make it clear who does what.
• Measure how many conversations end in a sale, not just how many messages you reply to.

When it makes sense to move to a tool with AI

If your business receives few messages per day and you can reply by hand, the basic WhatsApp Business automation is enough. You don't need more. However, if many messages come in, if you have several people handling them and they overlap, or if you notice you're losing sales because you don't reply quickly, then a tool is worth it. Wando is designed for LatAm SMBs, in Spanish and with local support, and it has an inbox with AI suggestions, templates, campaigns, contacts with labels, statistics, multi-agent, time-based auto-replies, and audio transcription. The trial is 14 days and doesn't require a card. If you want to see pricing, it's at wando.online/pricing.

You also have to be honest about the limits. AI doesn't replace the team's judgment: it suggests, and someone approves. If your business has very changing information or decisions that require human judgment, the approval flow is the right one. And if you've never loaded real responses, the system has nothing to learn from: learning starts when the team starts using it. The first few days the suggestions will be generic; with use they get adjusted.

What to look at before choosing an automation tool

Not all tools are the same, and not all of them work for you. Before deciding, review these points:

• That it uses Meta's official API. If they offer you an unofficial integration, the risk of being blocked is yours.
• That the model suggests replies with approval, rather than sending blind automatic responses, unless your case justifies it.
• That it learns from your real replies and not only from flows that have to be programmed and maintained.
• That it has a shared inbox if several people handle the same line.
• That you can see message and campaign statistics, not just a chat counter.
• That support is in your language and during your hours. If something breaks on a Friday afternoon, that matters.
• That you can try it before paying. Wando's trial is 14 days, no card required.

Frequently asked questions