The intent of a user is their purpose, the ultimate goal. Examples are ordering something, wanting to activate something on the user window, looking for shows, or simply saying goodbye. A chatbot should be able to perform some actions based on the intent it detects from the user's message.

Suppose we want to create a chatbot for a store that sells IT-related products. As a preliminary procedure, it is necessary to consider what actions the chatbot will be able to execute once requested by the user. For example, the chatbot will need to respond to the user with appropriate information when a user asks to see the products the store sells by supplying it: I want to buy a mouse. Similarly, when the user sends a message, such as Looking for a store in Rome, the chatbot should be able to locate all the stores near that particular location. To perform each of these actions, the chatbot must be able to distinguish between the two intentions of the user: searching for a product or a point of sale. In the following diagram, two possible intents are expressed by the user:

Detecting intents from the user's message is a very common problem in the field of machine learning. This is the technique called text classification, in which the objective of the program is to classify documents/phrases in several classes that represent the intent of the user. Understanding what the user's request is is the intelligent part of the chatbot because there are many ways in which a request can be expressed in natural language. The chatbot will try to interpret the user's request by identifying the closest intent. Naturally, the association will not always be precise; in fact, a ranking of possible interpretations is returned. But from this point of view the answers can be improved by providing more alternative examples of the same requests.