Microteaching AI Agent

The growing integration of Generative Artificial Intelligence (GenAI) — particularly large language models (LLMs) into education, is reshaping how teachers, students, and intelligent systems interact and assess learning. This transformation highlights the need for new assessment paradigms that can support dynamic, multimodal, and personalized learning experiences. Within this context, microteaching remains a foundational practice in preservice teacher development, yet its traditional evaluation methods often suffer from delayed, subjective, and narrow-scope feedback. To address these challenges, this study proposes a novel GenAI-driven, multimodal intelligent agent designed to evaluate microteaching performance. Built on the Coze platform, the system integrates audio, textual, and visual data to generate structured, individualized feedback. This approach empowers preservice teachers with timely, actionable insights that sup¬port adaptive professional development.