AbstractThe increased globalization of systems engineering is spawning a range of activities in the support of systems engineering teams who are geographically dispersed. The primary focus is on ways in which these team members can exchange information easily, quickly, and accurately. This paper takes the view that a network agent in a systems engineering design team comprises a computer plus a person, the two combining to form the intelligent network agent. A great deal of current activity is being put into the communication of data between computers. Relatively little is being put into supporting the communication between the people who are the most significant component of the “intelligent agent.” This article investigates the fundamental reasons why the communication is taking place and concludes that its purpose is to transfer from one agent to another sufficient information for the receiving agent to, eventually, supply a satisfactory deliverable to the sending agent. This implies the transfer, in some form, of a set of requirements. In such a transfer, these requirements will always be interface requirements. The paper goes on to discuss the transfer of computer‐based data between dissimilar tools and the issues involved in this transfer. It then compares the domains of “expertise” of the tool and of the human user, and discusses the mutually supportive operation of these together. From this discussion, the paper proposes a first‐cut set of requirements for the data exchange mechanism and its supporting environment. © 2001 John Wiley & Sons, Inc. Syst Eng 4: 213–221, 2001