Conversational Design in chatbots

Conversational Design: How to Apply to Your Chatbot

Conversational design is one of the most relevant points to capture the attention of your client. In this article, we tell you what it is, how to apply it, and some tools you can use. When you talk to someone, it’s very dull when they don’t interact much with you. It is also not pleasant when your tone is monotonous and uninteresting. It’s disheartening as well when the same happens with a chatbot – a technology that can be programmable. It can be unattractive and very damaging to the customer’s experience and involvement with their brand. It’s essential to keep the customer interested and make them want to talk to their virtual assistant about their product.

But what is Conversational Design?

According to Google, Conversational Design (or “Conversational UI”) is a type of language based on human conversation and experiences. The more natural and enjoyable the interactions, the fewer aversion customers will have to your chatbot. Nobody likes to ask a question and have a completely different answer, right? Well, the essence of Conversational Design is to have defined and logical conversation flows. That is, flows that resemble the reasoning and execution of a dialogue. Betting on Conversational Design is essential to build a consistent content strategy in your chatbot. The goal is to stimulate and keep the user interested in your product and your virtual assistant. Satisfy your clients’ requests and facilitate their curiosity. Indeed, you will have greater adherence.

Good Conversations Generate Good Business

The following situation already happened to you for sure. You go to a store, and the employee is unkind, or even rude to you. In the end, he may even have solved his problem, but the impression he left with was not satisfactory. What remains in people’s memories is the behavior someone has towards them. The way the company they go to treats their customers and what they do to keep them. In essence, whether the company applies a consumer-centered strategy or not. Customers want to feel consequential and special whenever they communicate with your company. So bet on providing them with a good conversation and, of course, solve their problems as well. All conversations you have with a customer are business opportunities. If your chatbot is well programmed and has good Conversational Design, your company will undoubtedly benefit from this investment.

5 Steps to Create Your Conversational Flows

Before forming your conversation flows, you must stipulate and follow some points. This way, you will have a consistent chatbot that will always be in tune with your brand.

1) Set Your Chatbot Goal

Know what you want to achieve with your chatbot. An assistant who doesn’t have well-defined goals is usually not successful. A chatbot can have many different functions. It can be programmable according to the purposes you are going to have, like customer support or product sales. They can also just be marketing chatbots for advertising campaigns or internal support within your company, such as for the HR department.

2) Identify Your Persona and Your Chatbot Tone

Knowing whom you’re addressing is a crucial point. If people don’t identify with your brand, they have no reason to get involved. Find out who your target audience is. Understand your age groups, and your professions. You can even be more thorough (if it makes sense for your brand) and know particular factors, such as preferred music styles. This data will be useful when you try to approach them and create some kind of relationship. From these parameters, you identify your chatbot’s tone when talking to your clients. If you are going to address younger people, maybe a less formal manner is the answer. If your persona is elderly, use a more formal tone. Or, why not, an informal and relaxed tone for older people? That’s what you’ll find out.

3) Choose Topics

It is essential to have the topics of your conversations well structured so that there is a line of reasoning. If your bot starts at one point and goes to a different one, the client gives up and will look for another means of contact that does not have the same problems. The client giving up is not what you want, because the preferred means of contact for users is by message. More specifically, chatbots, as shown in The 2018 State of Chatbots Report. Let us know what paths your chatbot can take a right from the start. Show the options you have and the services you offer. Use the hierarchies between categories and specify the themes for each subtitle. If you have a start menu, where one of the options is “Contract Changes”, the branches that follow you will be something like, “What data do you want to change?”

3) Write the Text

When writing the text, have this information in mind: for whom you are writing, what formality to use, and, of course, the product. Write according to the questions users ask or the doubts they have. Do not try to guess what the customer wants or needs. Follow the clues, and you will easily get your answers. You will write and create your products according to what the customer is looking for. Remember, the customer is at the center of decisions.

4) Manage Natural Language Processing (NLP)

It is crucial to have good knowledge of Natural Language Processing, especially in chatbots with artificial intelligence. NLP is the component that allows the bot to understand the “inputs” that clients write. Chatbots with NLP allow the user to have a normal conversation, providing a much more natural dialogue. On the other hand, in chatbots without NLP, free conversation is not possible. They only act according to scheduled interactions, where consumers can only choose the options. Your virtual assistant is intelligent, but it is up to you to increase your level of intelligence. Train him frequently and feed him with new ways to make an order. Only then can your “Machine Learning” algorithm improve.

5) Accept the Chatbot Mistakes

Keep in mind that chatbot is fallible. It’s a program, and yet it has its limitations. When you don’t have the answer, it’s better to say you don’t know than to try to answer and leave the customer dissatisfied. On those occasions, you can resort to suggestions, such as “did you mean X?”, or simply switch to a human assistant. These suggestions show the client that he is ready for any scenario and that he will do everything to serve it. Tools for Conversational Design

On the platform, you can choose your interactions with your client in a simple, fast, and intuitive way. That is, instead of just having monotonous answers, you can make the conversation more dynamic. In addition to simple text answers, you can choose from galleries, images, or even “gifs”. Any of these options makes communication with the user more stimulating and attractive. Consequently, it leads the customer to invest in your product and possibly become a regular customer. These enhancements are possible with’s back office. For this, you do not need to know how to program or have specific knowledge because it’s accessible to anyone.

Conversation Flows

The critical piece, however, is the conversation flow. On the platform, you have access to several blocks, each with its particular functions. You can streamline your interactions in the way that best suits your company and your use cases. To do so, you need to have your knowledge base (questions customers ask, expressions used in your sector, among others) already defined. This way, you can link the interactions of your chatbot with the information in your database. As you can see in GIF, there are four primary hypotheses of interaction with your client: Text, Image, Quick Reply, Galleries, and More. When the answer is straightforward and direct, you can use a block of text. For example: “Our store hours are from 9 am to 6 pm”. You can use the tab indicated for images if you need to “embellish” your conversation or simply complement your text block. You can easily copy and paste a picture of your choice. For a multi-purpose offer, you can use the quick reply option, the galleries, or a multi-purpose alignment. With these features, you give the customer the choice of the path they want.

More Advanced Options

In addition to these four response options, has more advanced options. With these, you can perform some actions that may be useful for your company. Imagine that the customer intends to make a simulation of the purchase of your product. To give you an answer that fits your profile, you need to collect your data. For that, you can use the “User Input” tool. This option allows you to collect the data provided by the customer and send him a personalized reply. In this case, with a simulation that fits the customer’s profile. This tool saves work to your customer support team, and the customer has an instant and satisfactory response. Another option that may be essential for your company is “Tags“. These allow you to count the number of people passing through a certain category. Knowing this information can be fundamental for improving and updating your product.

Integrations with API’s and Other External Systems also offers slightly more technical options, such as Webview or Integration Hub. These make it possible to integrate the conversation with external systems. The Webview allows the customer to view a web page within the chat. This functionality is interesting in some cases. For example, you need to collect user documents or request “login” to access information in your service’s customer area. Integration Hub is the feature that has won over our customers the most. Indeed your company already has several systems with which it works daily. With Integration Hub, you can integrate any system into your chatbot without spending too much time on your IT team. So you can send and receive data automatically. These tools allow the chatbot dialogue and service to become more complex and, therefore, more efficient. Having a well-structured and interactive Conversational Design is essential for your company to stand out in the market. With, you can achieve this and much more. Contact us, let’s talk!