Academic procrastination is a common behavior among students that negatively impacts their academic performance and overall well-being. It is associated with emotional factors such as anxiety, low motivation, and difficulties in self-regulation. Effective technological interventions aimed at addressing this behavior should not only assist with task management but also provide emotional support through active listening, emotion validation, and empathy. This work uses the Fred social affective robot programmed to assist students in organizing their activities, setting goals, generating empathetic reminders, and offering positive feedback. Fred integrates verbal communication with non-verbal affective behaviors to provide comprehensive emotional support. To assess user perception of Fred’s emotional awareness and support in addressing procrastination, we developed a set of interaction scenarios following a user-centered design methodology.. These scenarios simulated academic sessions where Fred accompanied students while doing homework. We conducted a between-subjects study with 73 high school students, comparing two experimental conditions: a neutral interaction mode versus an affective mode incorporating non-verbal affective behaviors, both using identical verbal content. The results indicate that the robot exhibiting non-verbal affective behaviors was perceived as significantly more pleasant and active compared to the neutral condition. These findings support the effectiveness of affective behaviors in enhancing user perception of emotional support, contributing to the design of emotionally intelligent conversational agents. Such agents have the potential for promoting emotional regulation and reducing procrastination in educational settings.