Complex products and their development processes may be viewed as systems, whose different aspects can be modelled as networks of interactions between elements in different domains. Many approaches have been proposed to explore, support or improve engineering processes by building such models. Developing these approaches, and applying them to problems of realistic complexity, often requires specialised computer software suitable for manipulating large data sets. However, creating suitable tools can be difficult–because software development is time-consuming and requires skills that many researchers and practitioners do not possess. We developed an approach which aims to address this problem by recognising the iterative nature of modelling research and its often tight coupling with prototype software development, and by reducing the effort of software prototyping and revision within this process. The approach is enabled by, and embodied in, the Cambridge Advanced Modeller (CAM)–a configurable software platform we have developed, refined and applied over several years and through a number of research projects.