Academic tutorials represent efficient methodologies for student follow-up, even more so if they are applied in an automated manner through the use of artificial intelligence algorithms. This study proposes an agile and low-cost computational solution for the management of academic tutoring through the application of the K-means algorithm. Using real student data from the State University of Bolívar, a segmentation was performed based on academic, socioeconomic, and territorial variables. The unsupervised approach made it possible to identify groups with common characteristics, facilitating institutional decision-making aimed at educational equity. The methodology included preprocessing of sensitive data, selection of the optimal number of clusters using the elbow method, and visualizations that reveal patterns in student performance and context. Additionally, an analysis of computational cost was carried out, demonstrating that K-means maintains linear complexity, achieving execution times under one second in most configurations. The results validate the use of this algorithm as an effective tool for educational environments that require scalable and efficient solutions. The study concludes that the integration of unsupervised algorithms with contextual analysis represents a robust strategy to strengthen academic support and tutoring programs, especially in public institutions with territorial diversity and vulnerable conditions.