In recent years, the global environmental science-policy interface has come to include a greater variety of knowledge. Social scientists have joined natural scientists at the policy table, and Indigenous and local knowledge is being taken ever more seriously. But this pluralisation raises political, normative, and epistemic challenges for environmental expert organisations, including with respect to how knowledge is managed, how it is judged to be valid, how it is made policy-relevant, and how it is presented to policy-makers and decision-takers. Based on an interview study of experts involved in the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES), we identify three broad approaches to these challenges: the integrationist logic, which seeks to combine all knowledge into a single ontology; the parallelist, which looks for similarities and connections between irreconcilable ontologies; and the pragmatist, which strives to apply knowledge when and where it will have the greatest positive impact. Rather than champion any one of these approaches, the paper explores their origins and how they negotiate paralyses to the timeliness of work. In avoiding ultimate formalisation of how value and knowledge pluralism are to be handled, IPBES allow more contextually sensitive practices to come to the fore. The paper concludes by discussing implications for environmental expertise more broadly.