The agriculture sector needs to do much better at innovation. Increased demand for food, plateauing productivity, changing patterns of competition and consumption, accelerating climate change, concerns over food safety, and the need to address these issues in a socially and environmentally sustainable manner, are just some of the reasons. These challenges are as relevant to Australia as they are to the world. To improve agricultural innovation, the broad prescription is that research and technology needs to be better coupled with market and policy changes that allow ideas and solutions to be deployed. Australian and international agriculture sector players, however, continue to grapple with questions on how to implement this prescription. In particular, on how to arrive at a mode of innovation that matches the ambition of transforming the performance and sustainability of the sector, both now and in the challenging years ahead. The purpose of this study was to develop a framework to better understand the relationship between different innovation configurations (partnerships, networks, and practices) and impact. Our starting assumption was that while configurations are contextually specific, broad patterns of practices and partnership associated with innovation and impact would emerge. Our logic was that these patterns could then form the basis of a framework to better explain how impact takes place, and point to tools and practices that increase the likelihood of innovation and impact. The study approach was to undertake theory-informed process analysis of the manner in which innovation and impact processes unfold over time. The key analytical perspective used was that of innovation systems, an empirically based concept underpinned by systems and evolutionary economics theories that explains the innovation process as a networked and socially embedded phenomenon, driven by evolutionary learning and systemic change.