Knowledge management is quickly becoming a requirement for today’s complex organizations. Creating and managing existing knowledge has been linked to successful innovation and to sustainable competitive advantage. However, systems specifically designed to manage knowledge, support knowledge creation, and verify existing knowledge are in their infancy. This article follows the framework for a Learning-Oriented Knowledge Management System, and shows how such a complex system can be supported by an equally complex technology – that is, a multi-agent system. We define single agents and multi-agent systems and subsystems in the context of knowledge management systems in general, and the Learning-Oriented Knowledge Management System (LOKMS) specifically. We show how a multi-agent system can be conceived to fully support the LOKMS, describe some necessary agents and agent subsystems, and demonstrate prototypically a multi-agent system designed and built to support the integrity-checking component of the LOKMS. This system begins the process of LOKMS design and development.