Land-use intensification is a major cause for the decline in species diversity in human-modified landscapes. The loss of functionally important species can reduce a variety of ecosystem functions, such as pollination and seed dispersal, but the intricate relationships between land-use intensity, biodiversity and ecosystem functioning are still contentious. Along a gradient from forest to intensively used farmland, we quantified bee species richness, visitation rates of bees and pollination success of wild cherry trees (Prunus avium). We analysed the effects of structural habitat diversity at a local scale and of the proportion of suitable habitat around each tree at a landscape scale. We compared these findings with those from previous studies of seed-dispersing birds and mammals in the same model system and along the same land-use gradient. Bee species richness and visitation rates were found to be highest in structurally simple habitats, whereas bird species richness--but not their visitation rates--were highest in structurally complex habitats. Mammal visitation rates were only influenced at the landscape scale. These results show that different functional groups of animals respond idiosyncratically to gradients in habitat and landscape structure. Despite strong effects on bees and birds, pollination success and bird seed removal did not differ along the land-use gradient at both spatial scales. These results suggest that mobile organisms, such as bees and birds, move over long distances in intensively used landscapes and thereby buffer pollination and seed-dispersal interactions. We conclude that measures of species richness and interaction frequencies are not sufficient on their own to understand the ultimate consequences of land-use intensification on ecosystem functioning.