Abstract:

(Systemic) Coaching as a solution and resource-oriented consulting method in work-world contexts is increasingly used in organizations to reflect on work and development processes. Coaching is also becoming more and more important in the field of academic education. Especially in subjects with high numbers of students, such as mechanical engineering, coaching has become an important measure to improve the quality of studies and to reduce dropout rates. The coaching market is becoming increasingly differentiated: in addition to classic face-to-face coaching and digital coaching (i.e. via telephone, email, chat or VR), coach bots or conversational agents are used, who train soft skills with employees in concrete work contexts and thus support their ability to reflect.

One of the central effectiveness factors in face-to-face coaching is the so-called joining or the development of a sustainable working alliance in the process – characterized by qualities such as empathy, appreciation and trust. Within the framework of the planned project, these factors are to be transferred to the coaching process between an AI coach or coach bot and a human coachee. It will be investigated which factors contribute to an effective working alliance in the human-machine coaching process.

The planned project is interdisciplinary at the interface of human-machine interaction and coaching effectiveness research. Within the framework of the full paper, the state of research will be surveyed. First, effectiveness factors from psychotherapy research as well as verbal and non-verbal coaching factors from positive psychology, solution-focused consulting approaches and systemic coaching will be compiled. Digital coaching provides complementary insights into joining factors. Bonding research on human-machine interaction has so far mainly focused on applications in customer communication or production. E.g. the way in which human-machine conversations are conducted, non-verbal aspects of joining/rapport or the empirical acceptance in human-machine communication are being examined.

In a second step, "key moments" for an effective working relationship in coaching will be operationalized in the form of a catalogue of questions on working alliance, instructions for action and intervention strategies, both from the research literature and exploratively from the authors' coaching practice. A coach bot avatar will be designed for these case studies with regard to verbal and non-verbal aspects and considering the operationalization developed. The experiment itself is planned within a wizard of oz setup. The full paper gives an outlook on the conception of such key moments of effectiveness in human-machine coaching. The use case and target group are engineering students of the Faculty of Process Engineering, Energy and Mechanical Systems at TH Köln, where about 1000 engineering students start their studies per year.

The planned project also includes the ethical dimension in AI coaching and explores the possibilities and limits of using AI-based systems in emotionally sensitive areas such as coaching. It asks where a human coach and coach bot can work together in hybrid teams, where coach bots can be more helpful than humans and where human coaches are irreplaceable. In this way, it contributes to research and development of conversation-based AI, to bonding research in human-machine interaction, to research on the effectiveness of coaching and to university didactics.

Keywords:

Coaching, coach bots, human-machine-interaction, digital transformation.