Summary * 1The spread and intensification of agriculture are recognized as two of the most important global threats to wildlife. There are clear links between agricultural change and declines in biodiversity across a wide range of agricultural systems, and convincing evidence that reversing these changes leads to a recovery in wildlife populations. * 2Nearly 4 billion euros are now paid annually through agri-environment schemes (AES) to farmers in Europe and North America to make environmental improvements to their land. Where appropriately designed and targeted, these schemes have proved successful in reversing declines in farmland wildlife populations. * 3We argue that insights gained from island biogeography and metapopulation theory, and from theoretical and empirical assessments of landscape connectivity suggest that AES may carry substantial wider benefits, which so far have not been considered in the design and deployment of such schemes. ‘Softening’ agricultural land could offset some of the negative impacts on biodiversity of the loss and fragmentation of non-agricultural habitats; could allow species to adapt to climate change; could slow the spread of alien and invasive species; and could contribute positively to the coherence of key biodiversity and protected area networks. Indeed, AES might represent the only viable way to counter these threats. * 4We outline a number of ways in which these wider benefits could be taken account of in the design of AES and suggest a number of characteristics of the species most likely to benefit from them. * 5Synthesis and applications. Agri-environment schemes might bring significant environmental benefits to habitats other than farmland by restoring the agricultural matrix that separates them. Theoretical and empirical research suggests that matrix restoration improves a number of ecosystem functions. Where they are available, AES might therefore represent a viable mechanism for addressing a range of pandemic environmental problems such as global climate change. Little consideration has so far been given to these wider conservation applications in the design, deployment and monitoring of AES.