Imagining Possibilities of an Artificial Intelligence Instructional Coach: Qualitative Evidence from District and Schools

• In Event: Exploring Teacher's Perceptions & Experiences With the Adoption of AI in Education

Abstract

This exploratory qualitative study examines veteran high‐school instructional coaches and teachers’ perceptions of integrating an AI instructional coach. Instructional coaching—a sustained, job‐embedded, one‐on‐one process shown to improve teaching—remains limited by high costs and scalability challenges. Guided by the Technology Acceptance Model, we conducted semi‐structured Zoom interviews with two coaches and four teachers and applied thematic coding in NVivo. Participants valued AI’s data‐analysis and real‐time feedback capabilities but underscored that AI cannot replace trust-based relationships and informal hallway interactions. Concerns centered on trust, contextual nuance, and feedback overload. No existing system offers a holistic AI coach spanning all coaching functions. Findings highlight both the promise and human‐centered gaps of AI coaching, informing future design and research.

Authors

• Yuxi Wen, Michigan State University

• Ao Qu, Massachusetts Institute of Technology

• Jiayi Zhang, National University of Singapore

• Alok Prakash, Massachusetts Institute of Technology

• Andrés Salazar-Gomez, Massachusetts Institute of Technology

• Paul Liang, Massachusetts Institute of Technology

• Jinhua Zhao, Massachusetts Institute of Technology