What is a flow chatbot

A flow chatbot is a program that responds to messages by following a predefined conversation tree: each option the user chooses takes them down a fixed path, like a phone menu but over chat. For example, if someone types "price", the bot shows numbered options and depending on the number they tap, it responds with pre-loaded text. It doesn't think or improvise: it executes the script someone programmed.

Why it matters to a small business

For a small business, a flow chatbot can be useful for very repetitive and closed-ended queries, such as hours, location, or order status. It's cheap to implement and gives immediate answers 24 hours a day. But it has a big limitation: if the customer asks something that wasn't in the script, the bot gets stuck or gives a generic response like "I didn't understand your question." And loading and maintaining those flows takes time: every change in hours, product, or promo requires editing the tree manually.

Concrete example

Imagine a clothing store that uses a flow chatbot. The customer types "Do you have shipping?" and the bot shows: 1) CABA, 2) GBA, 3) Interior. If they tap "1", it responds "we do same-day shipping." So far, so good. But if the customer types "Do you have the Beatles t-shirt in size M?", the bot doesn't know what to do, because that query isn't in the tree. The customer gets frustrated and ends up calling or going to the competition.

Common mistakes when using a flow chatbot

• Trying to cover every query with a tree: it becomes enormous, hard to maintain, and confusing for the user.
• Not updating the flows when data changes (hours, prices, stock) and the bot gives outdated information.
• Hiding the human contact: if the bot can't resolve it, it should hand off to a person quickly, not make the customer go in circles.
• Thinking the bot "learns on its own": a flow chatbot doesn't learn, it only executes what it was programmed to do.

How a flow-based chatbot is used in practice

A flow-based chatbot is set up from a platform like ManyChat or Chatfuel. The business owner builds a tree: writes a trigger (a word or phrase), defines the options the customer sees, and uploads the responses for each branch. Then they connect the tool to WhatsApp or Instagram and activate the bot.

Real example: a restaurant in Buenos Aires receives 50 messages a day asking about hours, address, and whether they take reservations. With a flow, the customer types "hours" and instantly receives the text with the week's schedule. That frees up the receptionist to handle orders or more complex inquiries.

The problem appears when the customer types "do you have a vegan menu?" and the bot doesn't have that branch. If there's no option to route to a person, the message is left hanging and the customer leaves. That's why any serious flow needs a clear exit to a human.

In Wando, on the other hand, the AI suggests responses to the team and the business approves them with one click. There's no need to program trees: the system learns from the real responses the team gives and improves on its own, always with human supervision.

What it's confused with and how it's different

The flow-based chatbot is confused with two things: the AI chatbot and the virtual assistant. The flow-based one is the oldest and most limited: it doesn't understand natural language, it only recognizes words or numbers and follows fixed paths.

The AI chatbot, like the one Wando uses, understands open-ended questions. You type "do you ship to Córdoba?" and the AI interprets the intent and suggests a response based on what the team has already answered. There's nothing to program: the system learns from the real history.

It is also confused with the virtual assistant, which is a program that executes tasks (scheduling, paying, checking stock). The flow chatbot only responds with text; it does not perform actions.

The practical difference: the flow-based one is a rigid menu; the AI one is an employee that learns. If your business receives varied inquiries, the flow-based one falls short and the AI one saves you hours of responding.

What happens if you ignore it

If your small business receives inquiries via WhatsApp and you don't use any bot, the cost is concrete: your team's time answering the same thing over and over. An aesthetics shop can spend 2 hours per day answering "how much does laser hair removal cost?" when that answer could be automated.

If you use a poorly configured flow chatbot, the cost is worse: customers who get frustrated because the bot doesn't understand their inquiry and there is no way out to a human. That translates into lost sales and a bad reputation by word of mouth.

The consequence of not using anything is that your team burns out with repetitive tasks and stops properly handling the inquiries that do require judgment. The consequence of using a poorly made flow is that the customer feels they are talking to a machine that doesn't listen to them.

The middle option, like Wando's, is for the AI to suggest responses and your team to approve them. That way you gain speed without losing control or the quality of service.

How it relates to Meta's official API

Meta's official WhatsApp API (WhatsApp Cloud API) is the legal way to connect a bot to WhatsApp Business. Without that API, any automation violates the terms of service and risks getting the number blocked.

A flow-based chatbot connects to Meta's API through an intermediary platform. That platform sends messages to Meta's webhook and receives the responses. Meta charges for messages separately, and the cost varies by country and volume.

Template approval (predefined messages for campaigns) depends on Meta and can take up to 48 hours. If your bot relies on templates to respond, you have to plan for that delay.

Wando uses Meta's official API, so there is no risk of getting blocked. And since the AI suggests responses in the moment, you don't rely on templates to answer common questions: responses are sent from the inbox with one click.

When a flow-based chatbot is NOT a good fit

A flow-based chatbot is not a good fit when your customers ask open-ended or varied questions. If you sell products with many models, sizes, or dynamic prices, the tree becomes huge and outdated. The customer writes "do you have the Nike sneaker in size 42?" and the bot doesn't know what to answer.

It's also not a good fit if your business depends on trust and personal attention, like a real estate agency or a clinic. A rigid flow can scare off a customer looking for advice.

And it's not a good fit if you don't have time to maintain it: every change of hours, promo, or product requires editing the tree. If you stop updating it, the bot gives out old information and damages your image.

In those cases, an AI solution that learns from your team's real responses (like Wando) is more flexible: there is nothing to program, and the system improves on its own with use, always with human approval.

Common mistakes of those who are just starting to understand it

The biggest mistake is thinking that a flow-based chatbot "learns." It doesn't learn: it executes what it was programmed to do. If you don't update the tree, it repeats old information and frustrates the customer.

Another mistake is trying to cover every inquiry with a tree. It ends up being a maze that nobody understands. Limit the flow to 5 or 10 closed questions and hand off the rest to a human.

It is also believed that the bot can replace a salesperson. It can't: it doesn't negotiate, it doesn't read tone, it doesn't build trust. The bot handles the simple things and the human handles the complex ones.

And a classic one: not testing the flow before activating it. Many businesses launch the bot and discover that the customer gets stuck in a branch with no way out. Test it with a friend, review the error messages, and make sure there is always an option to talk to a person.