Nowadays, Student Facial Expression Recognition Systems (SFER) are increasingly used in the context of smart classrooms, as they help educators assess their students’ emotional and engagement states during courses. However, the full potential of utilising the output of these systems has not yet been fully explored. In this chapter, we present a conceptual framework that explores how the output of an SFER system can be used to build a recommendation generator, a tool that helps instructors assess and respond to student engagement both during and after class sessions. We propose a system that integrates a Generative AI-based Recommendation engine designed to guide teachers in adjusting their teaching methods and improving overall classroom strategies. As part of this framework, we present the Smart Classroom Monitoring System (SCMS), a complete pipeline from data acquisition to dashboard visualisation alongside the AI Recommendation Generator Engine (AI-RGE). The AI-RGE, built using GPT-4o customisation using the OpenAI platform, generates recommendations for instructional actions based on students’ detected engagement states. We believe that this framework offers a promising advance for the education sector, providing valuable support for instructors seeking to optimise student engagement and learning outcomes. Future research is recommended to implement the proposed system in diverse classroom settings and further investigate its long-term impact on teaching effectiveness.