Types of WhatsApp Chatbots and Which One is Right for Your Use Case

Illustrator: Adan Augusto
types of whatsapp chatbots

The business opportunities arising from WhatsApp automation—since the first release of the WhatsApp Business App in January 2018 and WhatsApp Business API in August later that same year—have been getting more compelling. Each passing year brings new business feature updates from Meta as well as its official Business Solution Partners, adding more weight to the cause. One of the key reasons behind messaging automation becoming a major business-growth strategy, besides the growing consumer interest and demand, is chatbots. Virtual assistants, in combination with WhatsApp popularity among the general public, are truly changing the game. 

However, according to The WhatsApp-Led Growth Report, despite WhatsApp being a consumer favorite, only 30% of professionals currently use it. 

The reason? 

Most professionals, from marketing to customer support, struggle with the complexity of automation implementation as well as resource limitations. While automating with WhatsApp, like any other process, requires time and investment, the report showed that most perceived challenges and roadblocks stem from the lack of familiarity with available solutions that can ease the demands and streamline implementation.

Hence, this article offers a general overview of WhatsApp bot design and development, elaborating on the different technical options—types of WhatsApp chatbots available, and an analysis of their pros and cons. 

What is a WhatsApp Chatbot?

A WhatsApp chatbot is a conversational automation software developed to resolve friction points across the buyer journey via the WhatsApp channel. It offers an additional automation layer on top of WhatsApp’s outbound notification/campaign messages and quick replies. 

WhatsApp chatbots engage with users beyond a single message. They can work 24/7 and carry out multiple conversations with different people simultaneously. Their greatest advantage lies in their ability to serve a variety of use cases, automatically engaging with and answering customer questions regarding your products, services, or more.

Who Can Have a WhatsApp Chatbot?

To deploy a WhatsApp chatbot, you must first apply and be approved for WhatsApp Business API. WhatsApp values user safety and experience. Hence Meta ensures every business granted access to the API is in line with their Commerce Policy.

You can do so independently or via an official BSP. Managing the process via a trusted provider is more straightforward, with many offering additional features and services such as a campaigns manager, chat manager, chatbot builder, personalized customer support, etc. 

Types of WhatsApp Chatbots

Chatbot technology has undergone quite a transformation since its debut in the mid-90s. Admittedly the first AI assistants didn’t make much of an impression and gave bots a bad rep for decades to come. However, as far as technology goes, that’s ancient history. Conversational solutions today are not only more accessible but also helpful and reliable (with or without the use of AI). 

As things are, there are many different WhatsApp chatbot types and ways to develop them, which can be confusing if you are new to the topic. Neither of the bot types is necessarily better than the other. Which one is “right” depends on the use case, your company's needs, and preferences.

So let’s have a look at the options!

Rule-Based WhatsApp Chatbots

A rule-based chatbot works on the simple principle of a decision tree. This type of chatbot works by following a set of if-then (conditional logic) statements or rules created by human developers and responds to user inputs based on these rules. Decisions guide the user down a specific conversational path. There’s little space for deviation or unguided user input. 

Compared to the free conversation offered by an AI chatbot, this might feel significantly more restricted. However, the structure has its advantages since you can guarantee the experience they will deliver every time. In other words, rule-based bots are highly consistent and reliable, which is key to flawlessly functioning workflows. Despite not using AI, rule-based bots can power sophisticated workflows using a variety of integrations and powerups.

What are they best for?

Rule-based chatbots can be effective for handling simple and routine queries, where there are clear patterns and rules that can be followed, such as

  • Lead generation;
  • Lead qualification; 
  • Sign-ups, registrations, and bookings;
  • Account information updates;
  • Collecting feedback;
  • Quizzes;
  • Etc.

There are several different ways you can go about building rule-based bots from scratch, using a template or a text-to-bot generator.

From Scratch

First and foremost, you can develop the rule-based assistant from scratch. However, in this day and age, from scratch doesn’t necessarily mean from raw code. The market is flooded with low-code and no-code chatbot builders that enable the less-tech-savvy professionals to take bot building into their own hands. For instance, with Landbot, building a WhatsApp bot is a question of connecting blocks like it’s a game of LEGO. 

From a Template

Another way to go about it—especially if you are in a hurry to get a campaign rolling or are responding to a crisis—is by using a pre-designed WhatsApp chatbot template. Again, most bot development tools today feature template libraries with conversational flow solutions for any occasion. Landbot, too, offers the option to speed the development.


The last option is also the newest one. Thanks to the release of ChatGPT, our product team was able to experiment and come up with a quick yet extremely personalized way to build rule-based flows. They used GPT-3 model to create a chatbot that connects to the Landbot builder. This bot is able to build the flow (copy included) you need using solely your text-based prompt. 

It’s simple and quick, like a template, while offering a deeper level of personalization. 

AI WhatsApp Chatbots

An AI chatbot is a computer program designed to simulate human conversation using artificial intelligence techniques such as Natural Language Processing (NLP) and Machine Learning (ML). Lately, in relation to the ChatGPT release, you might have also heard the term Large Language Model (LLM).

To give you a bit of context… NLP is a subfield of AI that focuses on enabling computers to understand and process natural human language. It involves a variety of techniques, such as text analysis and speech recognition. Its primary goal is to create systems (assistants) that can understand human language in context, respond to it appropriately, and generate natural-sounding language.

Machine Learning (ML), on the other hand, is a broader field that encompasses a wide range of algorithms and techniques used to train machines to learn from data and make predictions or decisions based on that learning. ML algorithms can be used for a variety of tasks, such as classification, regression, clustering, and anomaly detection. ML models can be trained on a variety of data types, including structured data, images, and text.

LLMs like ChatGPT are types of machine learning models that are designed to process vast amounts of text data and generate natural language responses. These models are trained on massive datasets of text, allowing them to learn patterns and relationships in language that can be used to generate human-like responses. They use fancy deep-learning techniques like neural networks.

All in all, NLP is a subfield of AI that focuses on language processing and understanding, ML is a broader field that encompasses a wide range of algorithms and techniques used for machine learning and data analysis, and LLMs like ChatGPT are a type of machine learning model designed specifically for generating natural language responses.

Using an AI WhatsApp chatbot comes in handy in use cases when you need to handle more complex or unpredictable queries or when you need your bot to recognize context, not just input at its face value. 

For example, imagine a scenario where a hotel booking bot asks, “And how many people will be staying with us?” and the user responds, “just my husband and I”.  A rule-based bot, if not specifically trained to recognize these keywords, will report this as an error and bid the user to rephrases so it can process the information. It is most likely expecting the user to answer “2” or “two”.  On the other hand, a well-trained AI bot will be able to recognize the context and translate the user input correctly, responding, “Great, that will be 2 adults then.”

Now, let’s look at the different ways you can use AI in your WhatsApp chatbot. 

Using NLU Tools like Dialogflow

One of the ways to go is to take advantage of a natural language understanding platform like Dialogflow. Dialogflow is your typical NLU platform many companies use to design and integrate a conversational user interface into web applications, devices, mobile apps, interactive voice response systems, etc. Simply put, training an AI model is hard and time-consuming, so Dialogflow meets you half the way. It already boasts a pretty educated and well-trained database, so your developers just need to make a few tweaks. 

This type of bot is normally looser in structure and gives the user more control over the conversation, as user input is the main force driving it forward. However, this can sometimes backfire, leading to confusion and unpredictable user experiences. That's why, during NLP bot development, it's crucial to focus on conversation design to ensure a high-quality experience for the user. Think of it like this,  just like how a good script is key to a successful movie or TV show, good conversation design is essential for a top-notch AI chatbot experience!

You can either go full-Dialogflow or try to find the middle ground between the structure of rule-based bots and natural conversation. For example, Landbot no-code builder offers a Dialogflow integration that enables you to take customers through a structured conversation but defer to Dialogflow to offer a better experience once they select their use case. Not only do you get the “best of both worlds,” but you also ease the technical difficulty and likely lower the development time you would need if working with Dialogflow exclusively. 

For inspiration, visit our step-by-step WhastApp-Dialogflow chatbot tutorial!

Using Large Language Models (LLM) like GPT-3

Next, there’s, of course, the GPT-3 model by OpenAI, an artificial intelligence research laboratory. This LLM took the world by storm, and it’s still continuing to break status quo after status quo. 

GPT-3 stands for "Generative Pre-trained Transformer 3", which is, as mentioned, a powerful language generation model. It's actually one of the biggest and most advanced generative AI language models out there, with over 175 billion parameters. What's so impressive about GPT-3 is that it can generate written text that sounds just like something a person would write based on a prompt or input it is provided. It's been trained on a huge dataset of internet text and has the ability to come up with a wide range of outputs, from writing paragraphs to answering questions and even creating poetry and stories.

Due to its advanced capabilities, GPT-3 has caused a lot of excitement and interest worldwide—especially because it has so many potential applications, including WhatsApp automation. Still, it's crucial to note that, like all AI models, GPT-3 has limitations and biases that need to be considered when using it for various tasks. Hence, if you plan on using it for business, you can’t let it roam free without any direction. 

For instance, at Landbot, we have been focusing on finding ways to leverage the strengths of GPT-3, like its conversational aptness within controlled use cases such as FAQs. In the past, FAQ chatbots often ended up with a bad rep. Many times a bot would fail to understand the customer’s intention, providing a generic response that was not relevant. 

GPT-3, if applied correctly, can fix that. 

Now, using GPT-3 and Landbot, you can build a chatbot to answer FAQs and summarize answers, almost like a human. Clicking the link above will lead you to a tutorial that teaches how to build such a bot:

  • Using an existing database of common questions and answers, such as your FAQ page or Knowledge Base on your website, which serves as a source of information for the FAQ chatbot.
  • Configuring the GPT-3 integration inside Landbot to analyze the users’ questions and search the articles for the appropriate answer.
  • Creating friendly and clear responses on the go using GPT-3.

Feel free to try out a demo version of Landbot & GPT-3 bot to get a better idea of what you would be building. 

Built-in NLU Component

Last but not least, you can introduce AI capabilities into your WhatsApp bot by using a native NLU component by your chat builder of choice. Many low and no-code tools today offer their own AI features. 

At Landbot we are in the midst of developing an NLU block that will provide a layer of artificial intelligence to otherwise structured bots, allowing users to submit their answers in a way that feels natural to them, instead of following a rigid format. The feature is still in BETA but you can join the waitlist and be one of the testers to help us perfect this feature.

In Conclusion

WhatsApp chatbots offer another layer of automation to the popular messaging channel. Using an automated assistant is a great way to engage with your prospective and existing customers in a meaningful way without putting too much strain on your team. As we established in this article, there are different types of WhatsApp chatbots. Some are more complex than others but they all have something to offer if used in the right way and circumstances. All you need to do is ask yourself, what do I want my chatbots to accomplish?

Check out some creative WhatsApp chatbot examples here!