In this article, we propose an original method for providing fault tolerance in multi-agent systems through replication. Our method focuses on building an automatic, adaptive and predictive replication policy to solve the resource allocation problem of determining where agents must be replicated to minimize the impact of failures. This policy is determined by taking into account the criticality of the plans of the agents, which contain the collective and individual behaviors of the agents in the application. Some measurements assessing the efficiency of our approach and future directions are also presented.