This paper presents the architecture of an effective software application for managing enterprise projects. Viewing the execution of an enterprise project as a highly complex system in which many delicate trade-offs among completion time, cost, safety, and quality are required, the architecture has been designed based on the fact that any action in one part of such a project can highly impact its other parts. Highlighting the complexity of the system, and the way computational intelligence should be employed in making these trade-offs are the base of the presented architecture. The architecture is also based on the fact that developing a software application for appropriate managing of such trade-offs is not a trivial task, and a robust application for this purpose should be involved with an array of sophisticated optimization techniques. A multi-agent system (MAS), as a software application composed of multiple interacting modules, has been used as the main component of architecture. In this multi-agent system, modules interact with environment on-line, and resolve various resource conflicts which are complex and hard-to-resolve on daily basis. Based on the proposed architecture, the paper also provides a template software application in which an array of optimization techniques show how the necessary trade-offs can be made. The template is the result of the integration of several highly sophisticated recent procedures for single and multimode resource-constrained projects scheduling problems.