The present study applied Video Enhanced Reflective Practice (VERP) to promote and assess Finnish professionals’ interactional competence by means of stimulated recall during three group-based shared review discussions. A VERP trainer guided the professionals to reflect on children’s, their own, and colleagues’ actions and mental states based on video-recorded interactions in early childhood education classrooms. Self-reports on attuned interaction and guidance were collected before and after the in-service training. A coding scheme was developed to assess the target, type, and richness of mental state utterances (N = 787) produced during the VERP training. The professionals equally discussed children’s and adults’ mental states. They talked more frequently about motivational states when referring to children and cognitive states when referring to themselves and colleagues. The more attuned the professionals perceived their interaction with the children was before training, the more frequently they referred to their own or colleagues’ mental states observed during the group discussions. Our study extends assessment of mental states and provides guidelines for future group-based video-stimulated recall training for early childhood education professionals.