The Difficult Conversations Bot: Findings on Fostering Empathy and Reflective Communication Among Faculty and Staff

This paper presents final findings from the design and evaluation of the Difficult Conversations Bot, an AI-mediated role-play system developed to support empathetic communication among higher education faculty and staff. Grounded in Learning Engineering and informed by humanistic, social cognitive, and adult learning theories, the system was evaluated using a pre–post mixed-methods design. Quantitative survey results demonstrated consistent positive shifts in confidence, intentional perspective-taking, psychological safety, reflective practice, and perceptions of AI as a learning partner. Qualitative findings highlighted the value of practicing a difficult conversation in a psychologically safe rehearsal simulation with structured reflection. Together, these results suggest that AI-mediated role-play, when designed through a nested learning engineering cycle, can function as an adaptive, human-centered learning system for complex interpersonal skill development.