Abstract The transformation of natural habitats for farming is a major driver of tropical biodiversity loss. To mitigate impacts, two alternatives are promoted: intensifying agriculture to offset protected areas (land sparing) or integrating wildlife‐friendly habitats within farmland (land sharing). In the montane and dry tropics, phylogenetic and functional diversity, which underpin evolutionary values and the provision of ecosystem functioning and services, are best protected by land sparing. A key question is how these components of biodiversity are best conserved in the more stable environments of lowland moist tropical forests. Focusing on cattle farming within the Colombian Amazon, we investigated how the occupancy of 280 bird species varies between forest and pasture spanning gradients of wildlife‐friendly features. We then simulated scenarios of land‐sparing and land‐sharing farming to predict impacts on phylogenetic and functional diversity metrics. Predicted metrics differed marginally between forest and pasture. However, community assembly varied significantly. Wildlife‐friendly pastures were inadequate for most forest‐dependent species, while phylogenetic and functional diversity indices showed minimal variation across gradients of wildlife‐friendly features. Land sparing consistently retained higher levels of Faith's phylogenetic diversity (~30%), functional richness (~20%) and evolutionarily distinct lineages (~40%) than land sharing, and did so across a range of landscape sizes. Securing forest protection through land‐sparing practices remains superior for conserving overall community phylogenetic and functional diversity than land sharing. Synthesis and applications: To minimise the loss of avian phylogenetic diversity and functional traits from farming in the Amazon, it is imperative to protect large blocks of undisturbed and regenerating forests. The intensification required within existing farmlands to make space for spared lands while meeting agricultural demand needs to be sustainable, avoiding long‐term negative impacts on soil quality and other ecosystem services. Policies need to secure the delivery of both actions simultaneously.