The 5 C's of Chatbots (or How Chatbots Work)

We have already commented on another Article about chatbots What were they, what type of chatbots exist and What advantages and disadvantages do they have for companies. This time we are going to explain how chatbots work through their basic principles, what we call the 5 C's of Chatbots.

How Chatbots work through the 5 C's

1. C for Competencies: Chatbots have abilities

Competencies or intentions are those skills that we want to give to our bot. This will help us to know how far we want to go with the interaction between chatbot and user. To know what the competencies of our chatbot are, we must do the following questions:

• What is its generic purpose.
• What user problems we want to solve. Configuring a bot to solve specific data such as an address or a schedule is not the same as configuring more complex issues such as product types and variable prices.
• What specific list of functions it has. Listing them will help with the subsequent definition.
• What we don't want the chatbot to solve. It is important to be clear about this point, and to define it in order to act beforehand and avoid user frustration.

2. C for Cognitive: How Chatbots Understand

An AI bot will make use of its artificial intelligence skills to understand if the text entered by the user corresponds to one of the bot's competencies. Therefore, we must take into account a number of factors:

Semantic variations

How many different ways users can ask the same question.

Keywords

What keywords contain the phrase the user is making. Do an exhaustive investigation of those keywords that we must incorporate so that the bot can identify what type of question it is and has the resources to offer it.

Languages

What a predetermined list of words we have from the same competence, not only can they formulate it to us in different ways, but also in different languages (and language variations).

Not only does the bot have cognitive abilities applied to written text. It can also understand numbers, texts in an image, information in an image and video, identify gender and age of a face, understand the emotional level of a text and extract key words from a text.

3. C for Conversation: How chatbots communicate

To start a conversation for a user, it is necessary to devise the most efficient way to guide the user. We must first understand the context of the conversation, and then offer him the answer he needs. To achieve this, It is necessary to define:

• How many steps do we need to guide the user to their answer.
• What contexts to keep alive to affect the next interaction.
• What opportunities for dialogue do we detect for our marketing activities.

4. Channel C: Key to designing a chatbot

It is important to define the channel where the bot will interact with the user. Thus, a user who chats through an application, for example, Facebook or Telegram, is not in the same situation as a desktop user who interacts through a bot on a website. There are a lot of different channels, so we'll have to identify How do users of our channel behave.

5. C for Coaching: Chatbots are trained

A bot needs Constant attention so that we can get the best answer. Thanks to the demos and user tests, we will realize What do we need to improve until we get to the product we want. Key issues are:

• What questions are giving us the right answer.
• What types of questions have a low confidence index.
• What new questions the user has asked and which, therefore, we have discovered.
• What new integrations should we make.

These are the basic principles about chatbots. If there is a concept that has not been clear to you, we have a glossary of terminology about chatbots where you can answer your questions. If you think that maybe your company needs a chatbot, we have the knowledge to develop it.

How can we help you?

If you need more information, do not hesitate to contact us.