Carrying out distributed business processes over networks is rapidly shifting the nature of application architectures from the simple command and control client-server model to complex peer-to-peer models supporting dynamic patterns of social interaction and behavior among autonomous, proactive, goal oriented agents. Trusting agents to autonomously make decisions and execute actions on behalf of humans, as part of global business processes, requires both understanding and modeling of the social laws that govern collective behavior and a practically useful operationalization of the models into agent programming tools. In this article we present a solution to these problems based on a representation of obliged and forbidden behavior in an organizational framework, together with an inference method that also decides which obligations to break in conflicting situations. These are integrated into an operational, practically useful agent development language that covers the spectrum from the definition of organizations, roles, agents, obligations, goals, and conversations to inferring and executing coordinated agent behaviors in multiagent applications. The major strength of the approach is the way it supports coordination by exchanging constraints about obliged and forbidden behavior among agents. We illustrate this and the entire system with solution examples to the feature interaction problem in the telecommunications industry and to integrated supply chain management.