The societal value of strategies that delay the onset and progression of dementia cannot be overstated. Physical activity-unstructured and structured-is a promising, cost-effective strategy for the promotion of brain health. However, a large degree of variation exists in its efficacy. Therefore, to increase its utility as 'medication' for healthy cognitive ageing, it is imperative to identify key moderators and mediators of the positive effects of targeted exercise training on brain health. In this commentary, we focus on the type of targeted exercise training, the determinants of individual variation, including biological sex and genotypic factors, and the mechanisms by which exercise exerts its influence on the brain. We argue that a better understanding of these factors will allow for evidence-based, personalised, tailored exercise recommendations that go beyond the one-size-fits-all approach to successfully combat dementia.