There are three main ways in which intelligent agents can be used:

	The Human Surrogate that the capability of the intelligent agent to reason in a simple, rational manner. The agent works AUTONOMOUSLY without human direction, either in an actual or simulated (say, in a military simulation) environment.
	The Intelligent Assistant that supports the human in complex environments. The agent helps to unload more mundane tasks from the human. In this context, the agent, or agents, show autonomy, while they work cooperatively together with the human. In the first case the agents work independently of the human, then report back the results.
	An Architectural Paradigm for a software system that must integrate disparate subsystems. We could implement agents on one central computer, which would have total command at any time of all subsystems, and simply communicate in a hierarchical manner with the subsystems.