Within applied CA (Antaki 2011) workshops has been used as a way to show data to participants and make them reflect upon better professional practices (Perkins, Whitworth, & Lesser, 1997; Lock, Wilkinson, & Bryan, 2008; Antaki, Richardson, Stokoe, & Willott, 2015; Wilkinson, 2014). A constitutive feature of these workshops is to show data and recently more specifically: video data. Thus, the video clips are of utmost importance for the progressivity and learning potential in the workshop. However, the only principle for how to choose video clips for these workshops are that they should have a "trainable" potential. But what constitutes a trainable? This study focuses on how CA research can be used as a basis for skills development for professionals by discussing the concept 'trainable' in the CA literature. Using a research project as a case, the study discusses what might constitute and define possible trainables in video recorded data, and how CA researchers might facilitate learning activities for practitioners around these trainables. The data for the study is from an applied CA research project in a public service organisation. Through CA analyses of video recordings of video consultations between public service professionals and citizens, interactional trainables were identified. These trainables were presented to the professionals in workshops inspired by existing methods of feeding back results to practitioners within the field of applied CA (e.g. CARM (The Conversation Analytic Role-play Method; Stokoe 2011, 2014) and ViRTI (Videobased Reflection on Team Interaction; Due & Lange 2015). The workshops were video recorded and analysed with the aim of describing the workshop practices in more detail to promote critical reflection about the term "trainable" in the CA literature. We discuss whether the concept of trainable could be defined formally and how the facilitation of trainables might be developed further.