TEACHMate: Designing In-Context Instructor-Centered GenAI Support for Learning Management Systems

As GenAI becomes increasingly integrated into educational technologies, instructors face growing pressure to incorporate these tools into the preparatory work that supports teaching at scale. However, most existing GenAI innovations emphasize student-facing applications, offering limited support for how instructors prepare, adapt, and manage course materials within learning management systems (LMSs). To understand instructors' needs for GenAI-assisted preparation, we conducted a formative study with ten instructors and identified the key values they prioritize: personalization, transparency, and control. In response, we designed and implemented TEACHMate, a GenAI assistant embedded in the Canvas LMS that provides course-contextualized assistance, asks follow-up clarification questions, and gives step-by-step reasoning. An evaluation with twelve instructors showed that in-context support was valued for integrating GenAI into instructors’ existing workflows, conversational follow-ups helped instructors clarify intent, and step-by-step reasoning supported understanding and troubleshooting. Our findings offer empirical insights into instructor-facing GenAI design and point to opportunities for better supporting preparatory teaching work within scalable LMSs.