RESUMEN
This article investigates how a one-to-one support process based on the use of a whole dairy farm simulation tool helps both farmers to reflect on their production strategies and researchers to better understand the farmers' contexts of action and decision. The support process consists of a minimum of four discussion sessions with the farmer: designing the Initial Scenario and formulating a diagnosis, building and simulating the Project Scenario corresponding to the objective targeted by the farmer, building and comparing alternative scenarios proposed both by the farmer and the researcher, and evaluating the process with the farmer. The approach was tested with six smallholder farmers in Brazil. It is illustrated with the example of one farmer who aimed to develop his milk production by more than doubling his herd size on the same cultivated area. Two other examples illustrate the diversity of issues addressed with this approach. The first estimates the sensitivity of economic results to price variations of milk and concentrates. The second compares two scenarios in terms of forage supply autonomy. The discussion assesses the outcomes of the approach for farmers in terms of response to their specific issues and of knowledge acquired. The research outputs are discussed in terms of the value and limits of using simulation tools within both participatory action research and advisory processes.