Language is one of the processes that human beings use to be able to communicate effectively, this helps to share not only information, but elements of everyday life, to process all kinds of situations, ranging from instructions for a certain event to show and identify and express emotional states. With the passage of time when individuals have an inter-relationship with other people, through various elements, such as work and educational environments, the factors of collaboration and exchange of ideas become more important, since as more communication takes place, emotional processes can be identified, as these end up becoming factors. With the above, it can be assumed and indicated that emotional intelligence takes a greater relevance, since it contributes in a very important way to the way in which individuals process and manage emotions in various social contexts. In this article, a development using machine learning is presented, the main objective is to be able to have a simple and easy understanding, based on natural language. Contemplating a DataSet that contains conversational elements, for this study two strategies will be applied to classify and predict the intentions and emotions of users, the study seeks to determine which of these two approaches, have better performance to support the understanding of natural language. This type of research work, contemplating neural networks, is intended that the creation of tools such as conversational agents, can have a better response, be more accurate and supported by various techniques to be empathetic to user conversations.