New forms of field experimentation are currently emerging to support transitions towards sustainable agriculture, including “multi-actor experimental networks” (MAENs). Both in public policy and in academic research, such networks are increasingly presented as a promising approach for fostering sustainable farming system design. Many studies have inventoried, categorized and compared experimental processes to discuss them in relation to contemporary issues. However, to our knowledge, these studies have not considered how MAENs can be implemented, nor their various contributions to sustainable farming systems design. The present work therefore explores the mechanisms whereby MAENs, depending on the way they are managed, support participatory design processes. Drawing on concepts from the design sciences, we studied 11 MAENs established across Europe to support intercrop (IC) design for field crops. Data on the characteristics of these 11 MAENs and their contributions to IC design were collected through individual and group interviews with the network pilots, and the study of individual MAEN documents. The analysis provides three types of results. First, we identify nine generative functions, that is, various processes through which experiments contribute to IC design, including: (i) finding one best option or highlighting contrasts between different ICs; (ii) highlighting the conditions that must be met for an IC to achieve certain effects; (iii) discovering new ICs or properties of ICs; and (iv) supporting the emergence/continuation of collective action for IC design. Second, we highlight different ways to manage MAENs, in other words ways to manage several experiments (in space and time) with a view to supporting participatory IC design. We show that this involves (i) coordinating several objects under design within a network of experiments, (ii) managing the coexistence of experiments guided by different logics in the same geographical area, and (iii) developing interactions between the experiments at a given point in time and over time to support IC design. Third, based on the previous results, we show consistency between the various contributions of MAENs to IC design and the different ways in which the pilots managed them, and we highlight three strategies for managing MAENs to support IC design: MAENs supporting (i) R&D-led design; (ii) farmer-led design; and (iii) distributed design. All these results provide mechanisms, points of reference, MAEN types and characteristics to inspire and foster the reflexivity of R&D actors interested in developing such participatory networks in the future.