An Exploratory Study of a Generative AI Teacher Coach

Abstract

In low-resource educational environments, scalable teacher professional development remains a significant challenge. This exploratory study investigates the usability and perceived usefulness of a Generative AI (GenAI)–based teacher coaching platform designed to provide personalized, feedback-driven instructional support to teachers. The GenAI coach was built using established coaching and AI design models. The GenAI coach analysed audio from teacher talk during classroom teaching to provide personalised feedback and recommendations to teachers, including co-creating future goals and action plans. Teachers from a low-income K–12 school in Mumbai, India used the GenAI coach over multiple sessions. A mixed-methods evaluation included surveys and semi-structured interviews. Results indicated above-average usability and positive perceived value. Teachers highlighted strengths such as unbiased, personalised, evidence-based feedback, privacy and encouragement of self-reflection. Challenges included limited classroom capture, long processing times, and a need for lesson-specific feedback and pre-class planning support. The study offers early evidence that GenAI coaching tools can provide scalable and context-sensitive professional development in settings where access to human coaches is limited. Findings have implications for the future design of AI-enabled teacher development systems in India and other resource-constrained environments.