Meeting the challenges of current agriculture requires the development of innovative cropping systems with high environmental, economic and social performances. Agroecology seeks to optimize the services provided by cultivated and wild biodiversity to reduce the reliance on external inputs while maintaining/increasing cropping system sustainability. Maximizing the delivery of ecosystem services from biodiversity requires an in-depth redesign of production systems considering in-field management as well as ecological infrastructures surrounding the fields. An agroecological system experiment aims to design and assess cropping systems, the spatio-temporal arrangements of these cropping systems and ecological infrastructures at the farm level (Petit et al., 2021). Located on the INRAE Domaine d’Epoisses experimental farm (Bretenière, France), the CA-SYS platform is a collaborative research platform experimenting agroecology. Since 2018, an agroecological system experiment has been implemented on 125ha, and tests pesticide-free agroecological systems comprising four cropping systems and a high density of ecological infrastructures (Petit et al., 2021). The pesticide-free constraint is as well an objective to enhance the beneficial organisms involved in the biological regulation of pests and improve the environmental performance of the systems. The four cropping systems represent two main agricultural strategies that may foster biotic interactions: 1) two tillage-based systems (occasional ploughing, false seedbed, mechanical weeding); 2) two conservation tillage/agriculture systems (i.e. rotational/permanent no-till). These two options require mobilizing crop and cover crop diversity in time (at the scale of the crop succession) and space (e.g. mixture of crop species and/or varieties, companion crop). Legume crops play an important role in contributing to pest management (diversification of crop species and sowing dates) and in reducing N fertilizer. Other agroecological practices are combined in the cropping systems to have robust and competitive crops (e.g. choice of resistant cultivar, delayed sowing date, increased sowing density). The implementation of agroecological principles started in summer 2018 with the objective to reach similar profitability and productivity than in neighbouring conventional farmers at a 10-year horizon. Data are collected on a yearly basis to assess crop performances through agronomic diagnosis, systems performances through a multi-criteria assessment, and the ecological transition through a temporal analysis of ecological indicators. After three years of experiment, the performances of the four cropping systems were assessed with the MASC 2.0 multi-criteria assessment method (Craheix et al., 2012). The cropping114 systems have intermediate to high environmental performances (high air and water quality, biodiversity conservation, intermediate energy consumption, intermediate to low energetic efficiency) and intermediate social performances (high farmer’s working conditions but low supply in raw material). However, the cropping systems have low economic performances (low productivity and profitability). Yield losses due to difficulties to manage pest pressures (e.g. diseases on protein crops, insects on oilseed rape; slug and vole damages and weeds in no-till systems) partly explain the low economic profitability, despite the low input costs. Low profitability is also partly due to several crops, used to diversify crop rotations, but with low productivity and poorly valued in the current sectors. The difficulties in the management of some of these innovative agroecological cropping systems and their unsatisfactory performances have prompted some adjustment of farming practices. Further multicriteria assessment, i.e. at the end of the crop succession, will be necessary to analyse the impact of these adjustments. In addition, we expect in the coming years that the ecological infrastructures implemented in 2018 will increase beneficial organisms to improve pest management in the pesticide-free context of the platform. Methodological developments are needed to assess the whole agroecological system, including the ecological infrastructures.