Our planet is home to a rich diversity of arthropod vectors, vertebrate hosts, and pathogenic microbes, and while the zoonotic risk of some of these is well-known, novel zoonotic pathogens regularly spillover into humans. However, with such a rich diversity of vectors and vertebrate hosts it appears nearly impossible to forecast the pathways by which novel pathogens may spillover to cause zoonotic outbreaks. Using ecological interaction data and new models, we show how the complex (and seemingly chaotic) dynamics within vector-host communities can be quantified to identify high-risk vectors and reservoir hosts which may contribute to future zoonotic sillover events. Building on this, we show how this data can then be used to forecast the pathways by which known or novel vector-borne pathogens may spillover into human populations.