A Coaching Model that provides the basis for designing AI-based Coaching Conversations for Evidence-based Practice

Abstract

The aim of this research paper is to be able to design a coaching model that is generic, adaptive and flexible. The coaching practitioners either use general coaching models that are borrowed from other disciplines like psychology, philosophy, etc., or create their own coaching models that are domain specific. If these models are not tested, it leads to a perception that coaching is not supported by an evidence-base of underpinning coaching theories. For coaching to further evolve as a sound profession, there is a need to develop models for coaching that are supported by theoretical contexts and backgrounds relevant to coaching. Also, certain shared standards and codes need to be followed by the coaches. Further to this, as the use of conversational AI agents for coaching is becoming more widespread, the design of these systems needs to be guided by the same coaching theories, models and standards as well. Therefore, this study proposes a generic coaching model that can form the basis for designing AI-based coaching conversations. The coaching model suggests using GROW model for goal-focused coaching, and to provide a structure to the coaching session. The model also suggests integration of elements from solution-focused, cognitive-behavioural, and positive-psychology coaching to design conversational AI features based on theoretical underpinnings and standard coaching competencies.