Nowadays, AI is a real springboard for finding solutions to optimize and improve learning and teaching processes. This issue has been a focus of humanity for millennia, and very significant advances have been made in this quest. This article aims to address the issue of optimizing and improving learning and teaching processes through AI (Artificial Intelligence), considering crossroads of research fields AIED (Artificial Intelligence in Education), EDM (Educational Data Mining) and LA (Learning Analytic). The research made use of secondary data collected from previous research on the topic and primary data was collected using a case study. A comparative analysis was conducted and based on this opportunity, we propose a multi-agent system based on AI techniques, which is capable of performing broader analyses of learning and teaching processes. The research also implemented a prototype of EMAS. Through this system, teachers and learners will be able to access a wide range of relevant and reliable information about learning and teaching processes.