Since it is difficult (or even impossible) to assume anything about the agents' behavior and goals in an open multi-agent system, it is often suggested that an organization is imposed upon the agents, whichhich, by abstracting away from the agents, specifies boundaries and objectives that the agents, by enacting roles, are expected to adhere to. In practice, this is usually done by creating a middleware, which acts as a bridge between an organizational specification and the agents, often taking away too much of the agents' autonomy. This project investigates how to make agents organization-aware, thus removing the middleware and letting the agents directly reason about the organization. In this paper, we discuss the results so far, and describe the future goals and research direction for the project.