FaciliTrain: Practicing Facilitation Skills through AI-Simulated Group Dialogue

Facilitation is a critical skill for managing inclusive and constructive group dialogues, yet training is often resource-intensive, requiring human coaches and live practice partners. The authors introduce FaciliTrain, a voice-based AI system designed to scale facilitation training. The system simulates multi-participant conversations where users practice five evidence-based techniques—Validation, Invitation, Modeling Examples, Asking Follow-Up Questions, and Making Connections—and receive structured AI feedback to guide their reflection.

The researchers conducted a mixed-methods study with 24 participants, including a formative phase to refine the system and a controlled pilot (n=12) comparing an AI-feedback condition against a self-practice control group. Participants engaged in simulated group scenarios and were evaluated on their ability to apply the five techniques in a novel, un-scaffolded dialogue. The study utilized both quantitative performance metrics (F1 scores) and reflexive thematic analysis of participant interviews to understand the user experience.

• Externalizing Intuition: The system's taxonomy helped participants name and consciously apply facilitation behaviors they were already performing intuitively, making these skills more deliberate.
• Cognitive Demand: Participants identified 'Making Connections' as the most difficult technique, a finding supported by AI logs showing higher failure and retry rates for this specific task.
• Voice as a Forcing Function: Using voice rather than text prevented participants from over-editing their responses, forcing them to practice in-the-moment decision-making similar to real-world facilitation.
• Preference for Feedback: Despite comparable performance between the feedback and self-practice groups, participants overwhelmingly preferred the AI-feedback condition, viewing it as a necessary component of meaningful practice.