A randomised controlled comparison of the effectiveness of human and AI chatbot coaching with goal attainment, wellbeing and self-efficacy

Executive coaching is widely used as a tool to develop leaders and organisations. The profession is rapidly changing due to the introduction and explosion of AI chatbots, AI-enabled co-pilots and digital technology platforms. This raises the question of how AI coaching compares to human coaching. In this randomised controlled experiment with N = 114 coachees, accredited human coaches were directly compared with automated AI coaches for the first time. Outcomes were measured using established and validated progress psychometrics on coachees’ goals, motivation, resilience and wellbeing, and valid, momentary measurements. Our study emphasises a co-regulation model, which predicts better results for human coaches and minimal impact on attrition and dropout. The study demonstrated substantial effectiveness in diverse coaching outcomes but only for human coaching, with effect sizes in the mid-to-high range and strong predictions of success by initial self-efficacy and hope of the coachee. Implications and limitations of these findings for HRD theory and practice are discussed.