Multi-agent systems (MASs), comprised of autonomous entities with the aim to cooperate to reach a common goal, may be viewed as computational models of distributed complex systems such as organizations and institutions. There have been several model proposals in the agent literature with the aim to support, integrate, substitute human organizations, but no attempt has gone beyond the boundaries of this research context to become a mainstream software engineering implementation guideline, nor has it been adopted as a universal model of multi-agent interaction in economics or social sciences. In this work we counter top-down, operational organization specifications with a logical model of a fundamental concept: agreement, with the long-term aim to create a formal model of multi-agent organization that can serve as a universally accepted basis for implementation of collaborative distributed systems.