The integrated management of the biodiversity of our planet, sustainable development and social inclusion is one of the largest current challenges for the human kind. This management cannot be addressed only by technocratic means, as it would face the issue of acceptability and also miss potential solutions com-ing from the local practices and local stakeholders. Therefore, we need to explore and support participatory methodologies for combining as well as possible biodiversity conservation and social inclusion. Among these methodologies, companion modeling (ComMod, Barreteau et al. 2003) stands as one of the lead- ing reference. This approach combines: 1) a multi-agent-based modeling and simulation (MAS) of the underlying renewable natural processes (biological, ethological, physical. . . ); and 2) a social simulation of the decision processes between stakeholders, via some role-playing game (RPG), in order to provide the players of the game with some feedback on their decisions, individual and collective. In spite of all its merits and of being our main source of inspiration, this combined MAS/RPG methodology suf- fers from some limitations. At first, it presupposes a prior consensus of all players on a choice of a simulated model. Second, the covering by the simulation is incomplete and has some bounded temporal horizon, a consequence of its stochastic nature. We propose an alternative (and/or a complement) to MAS/RPG through a set of various assistants agents for helping players at individual decision and negotiation between them. Examples are: an agent proposing and capable of defending (through an argumentation process) decisions; an assistant agent for decision analysis (proximities of choices, dominance relation...); a mediation agent to propose compromises and their argumentation; and an assistant agent for assisting players in individual decision and also negotiation, based on a viability analysis (analyzing the consistency of one player objectives with its decisions and the dynamics of the eco-socio-system). In this paper, we will detail this viability-based assistant agent, compare it to the MAS/RPG approach, illustrate it by an example scenario, detail its design and way of use, and summarize some preliminary evaluation of use. We will conclude in analyzing future prospects.