AIS's method of solving problems is done by the sufficing reasoning cycle. This is the guarantee that information will cycle from Agenda Manager to Scheduler to Executor in time enough to ensure maximum latency between sensing and action.
To do this, the Agenda Manager tries to identify as many successive operations as possible, though not necessarily all. Rather than guarantee optimality, the Scheduler then tries to do the best it can with the (perhaps incomplete) agenda. Lastly, the Executor partially executes operations until a dynamic interruption occurs.
AIS is designed to work closely with a dynamic, complex, input-rich external environment.
Go to a discussion of this capability for multiple architectures.