Abstract: This study tests, empirically validates and refines the Tandem framework for co-designing climate services(Daniels et al., 2019, Daniels et al., 2020), to enhance its applicability and effectiveness. Intended as an inspirational guide for ‘good practice’, Tandem is practical and non-prescriptive and is designed to be tailored to context. We apply Tandem in three different geographic and socioeconomic settings: 1) a rural community in Indonesia, where smallholder farmers are confronting climate impacts on agriculture; 2) two cities in Sweden, where planners are addressing climate-related flooding and heat stress; and 3) communities and institutions in a Colombian river basin, where climate change is leading to water scarcity, raising questions about equitable use. We find that Tandem was effective in these settings in: 1) moving from ‘useful’ to ‘usable’ information by building trust; 2) increasing institutional embedding through strengthened relationships and networks; 3) improving climate information uptake and use; 4) increasing capacity, confidence and a shared understanding of climate information by users, and the decision context by providers; and, 5) serving as a non-prescriptive guide for users, intermediaries and providers to co-design and structure an effective process for collaborative learning and action. We use insights from these case studies to enhance the original framework, enabling it to 1) scope and review climate and non-climate vulnerability and risks; 2) incorporate gender, social equity and power considerations; 3) acknowledge the value of local and traditional ecological knowledge; 4) co-explore horizontal and vertical governance at appropriate decision-making scales; and, 5) provide flexible starting points, with early identification of impact indicators.