Basic concepts to create a chatbot with artificial intelligence.
The world of chatbots is in full boil. If you are here, it is that you have noticed it. However, the amount of information that is emerging in this regard causes that in many occasions they end up mixing "churras with merinas" and therefore you do not finish to find out 100% of what they are talking about is a "chatbot", a " platform to build it "or an" artificial intelligence tool ".
When creating a chatbot with Artificial Intelligence, there are three concepts that you must control:
Intentions (#intent).
Entities (@entities).
Dialogs (Dialog).
It should be noted that, although each of the tools of Artificial Intelligence (or NLP, natural language processing) on the market can call these concepts in another way, we will take this as the standard nomenclature.
Intents: the Artificial Intelligence system, the first thing it tries to identify when someone writes to it is the "intention" of what the user wanted to say. For example, if we are in front of a chatbot to reserve tickets, we will have defined different intentions of the business such as # BuyEntrance, #ModifyReserve, # ReturnInput. Where does that # come from? Simply to differentiate intents (# intentions) from the following concept that we will see, the entities (@entidades) that go with the @.
Entities: are those user inputs (words, categories, phrases) that will determine the response of the chatbot, since they are necessary to carry out the action. Entities are associated with intentions. For example, imagine that our service allows you to buy tickets for theaters and concerts, so for the intent #CompraEntrada, we can define the entity @TipoEntrada that will have two types of values (theater and concert) as well as @FechaEntrada and @HorarioEntrada in case more than one schedule for that event, and @CiudadEntrada.
Dialogs: the dialogue is the structure of the conversation itself. In dialogs, we define what we want the chatbot to say when it identifies what the user wanted to say (ie, the intent). In this way, and visually, a conversation is created in tree mode, where we will indicate what the chatbot should answer when it recognizes the intention that the user wanted to say. That is, if the user says "I want to buy an entry", the IA engine (Watson for example) will identify the intent # BuyEntrance, and at that point we will have configured the dialogue so that the IA engine says: "Great, I'm here to help you, did you want a ticket to a theater or a concert? "; and so on.
Source
https://planetachatbot.com/conceptos-clave-para-construir-chatbots-fe67d977dfd