This research defines the business requirements for forming a Network capable of instantiating Virtual Enterprises (VEs); investigates the current state of the art for enterprise and Information and Communication Technology (ICT) architectures; and develops an improved intelligent software architecture for managing a dynamic, VE-forming Network of Partners. What is lacking in the current body of knowledge are methodologies and reusable partial models for how to specify and design ICT architectures that support complex coordination in an agile and intelligent manner. There is no generic method or comprehensive architecture today which would give businesses the ability to build and maintain such collaboration in a dynamically changing virtual enterprise. This problem is particularly evident in one-of-a-kind production, because the VEs must be created every time a large project is undertaken. Enterprises need to remain competitive by developing products which are specific to individual requirements, and developed globally to leverage resources. This research proposes that an enterprise can be represented as a set of hybrid, flexible, autonomous agents which encapsulate decision centres and other resources in order to execute business functions. The goal of this research is to define and develop an improved software architecture which supports globally distributed functional and resource management. The proposed ICT architecture will be developed and demonstrated using state-of-the-art protocols and platforms. The focus is on supporting: intelligent and agile interaction with customers and other Network stakeholders optimisation of the supply chain through collaboration and reasoning integrating heterogenous information and human activity systems The research objectives are to answer the following questions: 1. What are the business requirements, design parameters and design constraints for managing a dynamic, VE-forming Network of Partners? 2. Based on this design space, how can suitable supporting software architectures be specified? 3. Is the developed hypothesis valid? The fundamental research hypotheses are: 1. a VE-forming Network of Partners can be designed as a multiagent system of hybrid agents 2. such a design can be implemented using current protocols and platforms The research specifies and validates the architectural design of the Agile-Oriented Architecture, which contains several important design attributes currently not present in any of the current ICT architectures.