Purpose
Feedback is essential in medical education, yet many faculty lack training and confidence in delivering it effectively. Traditional development methods are valuable but often resource-heavy and hard to scale. To address this, we created a blended faculty development model combining lecture material with interactive, case-based practice using AI-powered avatars. This approach offers a scalable, efficient, and realistic way to build feedback skills.


Methods
Nine interprofessional faculty – ranging from assistant to full professors – participated in a one-hour session that began with a 20-minute interactive lecture on the “ask-tell-ask" feedback model. Faculty then participated in small, facilitated groups in which they were each tasked with delivering feedback to a student avatar across one of three scenarios.
After each scenario, participants self-reflected, used the rubric to self-assess, and received feedback delivery ratings from the AI avatar. Group debrief discussions followed each scenario to reinforce learning. A 14-item evaluation survey captured participants' perceptions across key domains, including engagement, realism, learning, and future use. Open-ended responses were also collected.

Results
All participants completed the evaluation. They unanimously agreed that the session was effective, engaging, realistic, and enhanced their ability to give feedback. The majority reported that the AI avatars felt realistic and preferred this method over traditional role-play. Notably, no participants disagreed with any survey item, and all indicated they would recommend using AI-Avatars in future feedback training sessions.

Conclusion
This pilot highlights the potential of AI-powered avatars to enable scalable, effective faculty development in feedback delivery. Early findings suggest strong engagement, perceived realism, and educational value. Future iterations will explore asynchronous formats, broader implementation, and continued technological refinement.