Augmented reality has been broadly employed to help remote individuals communicate and coordinate. In this study, we develop and test a model that explains how augmented reality can facilitate collaborative diagnosis on an unexpected technical breakdown involving two complete strangers. Drawing on the affordance theory, we integrate the dual-task interference literature to reveal frustration valence and arousal as the underlying mechanisms. We tested our hypothesis in a laboratory experiment involving a custom-built augmented reality environment and physiological measurements. Overall, this study contributes to information system literature, human-computer interaction literature, and dual-task interference research by unearthing the effects of augmented reality characteristics on enhancing collaborative diagnosis performance.