Over the last few years, the use of agent-based simulations to study social systems has spread to manydomains (e.g., geography, ecology, sociology, economy). These simulations aim to reproduce real life situationsinvolving human beings and thus need to integrate complex agents to match the behavior of the simulatedpeople. Therefore, notions such as cognition, emotions, personality, social relationships or norms have to betaken into account, but there is currently no agent architecture that could incorporate all these features andbe used by the majority of modelers, including those with low levels of skills in programming. In this paper,the BEN (Behavior with Emotions and Norms) architecture is introduced to tackle this issue. It is a modulararchitecture based on the BDI model of cognition and featuring modules to add emotions, emotional contagion,personality, social relationships and norms to agent behavior. This architecture is integrated into the GAMAsimulation platform. An application of BEN to the simulation of the evacuation of a nightclub on fire is presentedand shows the complexity of behaviors that may be developed with this architecture to create credible andexpressive simulations.