With the advent of the Internet and the progressive development and consolidation of a wide range of web standards and technologies as well as the advances in distributed artificial intelligence (DAI), namely the multi agent system concept, new opportunities have emerged for conceiving, modelling and enhancing shop floor's performance and response. Modern IT-supported production paradigms denote a common concept where the shop floor is a lively entity composed by interacting intelligent modules whose individual and collective function adapts and evolves ensuring the fitness and adequacy of the organization, owning the system, in tackling profitable but volatile business opportunities. The self-organizing and peer to peer nature of these systems renders a collective behaviour and dynamics that are fundamentally new. Conventional diagnostic methods and tools have not been designed targeting the envisioned systems therefore lack the required support. In this paper the emerging IT-based production paradigms are surveyed as well as the existing diagnostic methods whose adequacy is analysed. The resulting requirements and characteristics are exposed to stress the need for rethinking current diagnostic practices in future automation systems. (C) 2011 Elsevier B.V. All rights reserved.