Easing the cognitive load of general practitioners: AI design principles for future-ready healthcare

Abstract

General practitioners (GPs) worldwide face increasing cognitive demands, especially in after-hours and voluntary primary care, where urgent decision-making and resource constraints exacerbate workload pressures. Studies across North America, Europe, and Asia indicate that GPs encounter similar challenges globally, with administrative burdens and patient complexity contributing to high cognitive loads. While prior research has examined technological interventions, workflow optimization, and cognitive assistance independently, an integrated, actionable framework tailored to GPs’ needs remains lacking. This study employs a design science approach to develop and evaluate a Neural Assistant for Optimized Medical Interactions (NAOMI), a prototype AI agent designed to support triage and clinical decision-making in after-hours and voluntary care settings. Through 80 simulated consultations and clinician feedback, we identify three key design principles: Comprehensive Data Collection and Analysis, Clinical Reasoning Transparency, and Adaptive Triage and Risk Assessment. These design principles provide a structured foundation for developing AI-driven solutions that reduce cognitive burden, enhance clinical workflows, and improve healthcare equity. By advancing AI integration in primary care, this study offers a scalable roadmap for AI-driven healthcare research and innovation, addressing systemic workforce challenges while optimizing patient outcomes.