This paper gives an example of how predictive models of the deterioration of reinforced concrete pipes and the consequences of failure can be used to develop risk-based replacement strategies for redundant reinforced concrete pipe networks. It also shows how an accurate deterioration prediction can lead to a reduction of agency costs, and illustrates the limitation of the incremental intervention step algorithm. The main conclusion is that the use of predictive models, such as those developed by Oxand S.A., in the determination of replacement strategies for redundant reinforced concrete pipe networks can lead to a significant reduction in overall costs for the owner of the structure.