Agents that live in robots: how are successful applications built?

Agent technology has today a wide spectrum of possible applications. It is easy to notice many examples of agents in different branches: economics, management, manufacturing, telecommunication, robotics, and entertainment to list a few. However, this technology is not mature enough yet to give precise predictions about success or failure of an agent-based project. This paper tries to demonstrate the current state of the agent technology. We describe important features of existing multi-agent paradigms to start some discussion on their suitability and applicability in robotic applications.