This research-in-progress paper focuses on examining configurations of collaboration between physicians and AI in decision making. From prior literature, we know that complex decisions in hospitals are the result of a collaborative decision-making process among physicians in a team. However, research from an information systems (IS) perspective in this area has so far focused on individual’s interactions with AI, while collaboration in decision-making for complex clinical cases reflects common practice also in technologically supported environments. Therefore, we aim to shed light on the question “which configuration of human AI collaboration in decision making is most recommendable for AI-enabled systems?” We plan to conduct a scenario-based experiment to investigate accuracy, speed, and satisfaction with various configurations of physician AI collaboration in the context of computer-aided intelligent diagnosis (CAID) systems. Our primary contribution will be a multidimensional evaluation of selected collaboration configurations aimed at improving healthcare with technology.