The working relationship between the military, U.S. government agencies, and non-government aid organizations can be a difficult one. The differing cultures of each organization have evolved out of contrasting missions and activities, resulting in different values, modes of interpersonal interaction, and approaches to work. Multi-cultural collectives have known coordination problems (e.g., Burke, Hess, Priest, Rosen, Salas, Paley, et al., 2005), but most efforts to enhance cross-cultural coordination do not take a comprehensive approach that develops individual and collective knowledge and behavior (Roberson, Kulik, & Pepper, 2003). Our team researched the requirements for developing interagency coordination at the field level during stability, security, transition, and reconstruction (SSTR) operations. This paper describes our findings and their implications for designing a computer-based interagency planning environment. We found that conceptualizing interagency collectives as multi-team systems (Mathieu, Marks, & Zaccaro, 2001) provided a theory-based method for identifying what must be developed in order to achieve successful interagency coordination. The multi-team transition (or planning) phase may be characterized as an interests-based, multi-party negotiation -- a collaborative problem solving task in which innovative solutions are sought through consensus building. The success of multi-team planning or interagency consensus building, in turn, is mediated by general strategies for success, including interpersonal relationship building and cross-cultural communication. We determined that the capabilities of web-based knowledge management systems and latent semantic analysis, an automated text analysis technique, can be integrated into a comprehensive training system that addresses individual and collective knowledge and behavior.