Animal movements are important drivers of nutrient redistribution that can affect primary productivity and biodiversity across various spatial scales. Recent work indicates that incorporating these movements into ecosystem models can enhance our ability to predict the spatio-temporal distribution of nutrients. However, the role of animal behaviour in animal-mediated nutrient transport (i.e. active subsidies) remains under-explored. Here we review the current literature on active subsidies to show how the behaviour of active subsidy agents makes them both ecologically important and qualitatively distinct from abiotic processes (i.e. passive subsidies). We first propose that animal movement patterns can create similar ecological effects (i.e. press and pulse disturbances) in recipient ecosystems, which can be equal in magnitude to or greater than those of passive subsidies. We then highlight three key behavioural features distinguishing active subsidies. First, organisms can transport nutrients counter-directionally to abiotic forces and potential energy gradients (e.g. upstream). Second, unlike passive subsidies, organisms respond to the patterns of nutrients that they generate. Third, animal agents interact with each other. The latter two features can form positive- or negative-feedback loops, creating patterns in space or time that can reinforce nutrient hotspots in places of mass aggregations and/or create lasting impacts within ecosystems. Because human-driven changes can affect both the space-use of active subsidy species and their composition at both population (i.e. individual variation) and community levels (i.e. species interactions), predicting patterns in nutrient flows under future modified environmental conditions depends on understanding the behavioural mechanisms that underlie active subsidies and variation among agents' contributions. We conclude by advocating for the integration of animal behaviour, animal movement data, and individual variation into future conservation efforts in order to provide more accurate and realistic assessments of changing ecosystem function.