Modern agriculture, especially monocultures, is increasingly blamed by proponents of agroecology for loss of biodiversity in agroecosystems through loss of natural vegetation and crop genetic diversity. However, loss of natural vegetation to agriculture has slowed in recent years, and the claim for 75% loss of crop genetic diversity due to the widespread use of improved crop varieties lacks evidence. The main objective of this perspective paper is to identify and analyze the kinds of diversity already developed by agricultural research in the context of their function, need and relevance to particular agroecological elements and principles related to biodiversity. The paper shows that biodiversity is already incorporated into agroecosystems in a multitude of ways such as improved crop varieties with functional traits; crop mixtures and dual-purpose crops in monocultures; inter-crops, strip crops or relay crops; cover crops; crop rotations; field margins; landscape diversification; crop-livestock systems; and agroforestry. Emphasis is also given to crop-associated biodiversity above ground and soil biodiversity. The paper highlights that more research is needed to (a) breed crops for multi-cropping systems; (b) improve understanding of how components of diversity interact within or across systems and scales; (c) understand the role of soil biodiversity in soil function; and (d) successfully use crop-associated biodiversity to manage pests and beneficial organisms. In the future, the promotion of agroecological principles should incorporate inputs from crop breeders, agronomists, ecologists and crop protectionists for the benefit of farmers, or such principles will remain aspirational.