Designing Synthetic Work Relationships: An Exploratory Study of Agentic AI Roles, Cognitive Load, and Psychological Need Satisfaction

There is no turning back from the adoption of artificial intelligence (AI) in the workplace. As AI emerges as an agentic partner in organizational contexts, it reshapes traditional collaborative structures. The central question is no longer whether AI will be used, but how such systems are integrated into complex knowledge work, how individuals navigate human-AI partnerships, and what consequences arise for human workers. However, the design of these partnerships affects not only task performance and organizational control but also the psychological work experience, particularly cognitive load and motivational engagement. In this study, we use a qualitative design in which participants modeled business processes while collaborating with AI partners assigned the roles of assistant, colleague, or supervisor. Applying Cognitive Load Theory and Self-Determination Theory, we examine how these role configurations shape task performance, mental effort, and the satisfaction of psychological needs, namely autonomy, competence, and relatedness. Our findings highlight trade-offs in human-AI collaboration and offer both theoretical contributions and practical design implications for configuring synthetic work relationships – defined as continuous, role-based socio-technical associations between humans and AI – that foster sustainable coexistence by safeguarding human agency and cognitive well-being.