The complex nature and multilevel scale of the challenges faced by agricultural research for development (AR4D) call for appropriate design and evaluation methods. At the design stage, this implies thinking about how new agricultural technologies enable certain changes within desirable future scenarios, and the systemic transformations needed for these technologies to have impacts. We have developed a method, called 'Impact Weaving', which combines participatory foresight and quantitative modelling to design more plausible anticipated impact pathways of the use of new technologies. Two case studies are used to present the method. We argue that time, skills and investment capacity are required to enable collective sense-making and to estimate the potential users of the technology. Nonetheless, applying this transdisciplinary approach at the early design stage will lead to more grounded technology dissemination that accounts for the constraints and aspirations of its users.