Abstract This paper discusses the role of complex systems models as both embodiments and sources of knowledge in the context of strategic decision-making. Models are frequently seen as one way of gaining knowledge of the future. Advances in the simulation of complex adaptive systems enable models to increase knowledge of system behaviour, but may also indicate the limits of such knowledge. The relationship between models and knowledge is discussed in the context of a simulation of the telecommunications industry. It is shown that the ‘knowledge’ provided by models is multi-valued and highly dependent on context; not single-valued answers but rather a statement of options, which place limits on the extent to which control can be exercised. This pushes much of the decision-making back to higher cognitive levels where objectives and values dominate. By high-lighting the potential trade-offs in any decision, complexity science can ensure a real debate about what values should be, or are being, adopted by society.