Emotions are part of human life, and they are present on several occasions, like decision making and in social interactions. Computational identification of emotions in texts can be useful in many applications, especially in distance learning courses. This research introduces an animated pedagogic agent, integrated to a Moodle virtual learning environment, with the objective of assisting the tutor in accompanying students, helping the students to acquire knowledge, identifying their emotions, and motivating the student to participate in activities and discussions. As a way of assessing students' emotional state, an experiment was conducted using real data from a completed course, involving students. The results obtained are promising, evidencing the importance of knowing the emotional state of the students, contributing to the learning process.