Teams are increasingly asked to solve complex and novel problems. From a broad range of domains such as the military, healthcare, and industry, difficult problems requiring the adaptation of a diverse set of expertise to one-of-a-kind situations are becoming more commonplace. These types of performance contexts require collaborative problem solving; however, the bulk of research on teamwork has dealt with behavioral coordination in routine tasks. This leaves a gap in the theory available for guiding design and training interventions to support collaborative problem solving, or knowledge-work, in teams. This paper addresses this gap by 1) providing a review of relevant theoretical issues, specifically the team cognition and externalized cognition literature, 2) advancing a theoretical framework and propositions rooted in this literature that relate the role of group process and external representations of problem spaces on problem solving outcomes, and 3) discussing future directions for testing, applying, and refining this model.