The provision of floral resources for the enhancement of beneficial insect populations has shown promise as a strategy to enhance biological control and pollination in agroecosystems. One approach involves the provision of a single flower species while a second involves the multiple flower species, but the two have never been compared experimentally. Here we examine the influence of single and multiple species flower treatments on the abundance and foraging behaviour of key beneficial insects in two agricultural agroecosystems (broccoli and lucerne crops). The five flower treatments comprised buckwheat only, phacelia only, a simple mixture of buckwheat and phacelia, a complex mixture of buckwheat, phacelia and a commercial seed blend or the existing crop as a control. The abundance of bumble-bees (Bombus hortorum) and honey bees (Apis mellifera) was highest in the three treatments that contained phacelia, while hoverfly (Melanostoma fasciatum) numbers were high in all four flower treatments. Bumble-bees and honey bees probed almost exclusively phacelia flowers, even when provided with a choice of other flower species in the simple and complex mixture treatments. In contrast, hoverflies probed the flowers of all plant species in single and multiple species treatments, with no apparent difference in acceptance. However, in mixture treatments, the majority of individual bumble-bees, honey bees and hoverflies probed the flowers from only one species, despite the presence of alternative flower species. Our results illustrate how an appreciation of insect floral attractiveness can be used to customise the species composition of floral patches to potentially maximise biological control and pollination in targeted agroecosystems.