Abstract
This developmental paper proposes a within-person model examining how daily AI deferral- the act of delegating task steps or judgments to AI- shapes employee functioning through two simultaneous pathways. Drawing on Job Demands-Resources (JD-R) theory, we argue that AI deferral can increase job resource gains (task control, role clarity, and time/bandwidth) while also reducing social support availability, particularly in interdependent work contexts. These pathways are expected to shape next-day adaptive performance and coping capacity. A key conceptual contribution is the distinction between AI deferral, an agentic input behaviour, and social substitution, a downstream relational consequence construct that is often conflated in existing work. The paper also outlines a feasible empirical strategy combining a 7-workday Prolific diary study with a randomised experiment to test causal claims. Feedback is sought on construct boundaries, measurement, and research design before full empirical development proceeds.