A knowledge management system must support the integration of information from disparate sources, wherein a decision maker manipulates information that someone else conceptualised and represented. So the system must minimise ambiguity and imprecision in interpreting shared information. This can be achieved by representing the shared information using ontologies. There are typically two approaches to developing ontologies to support decision making. In one approach, ontologies are developed to support new business processes or decisions, but often are not built from existing data repositories. In the other approach, ontologies are developed from existing data repositories, but often may not support new business processes and decisions. In this paper, a methodology for knowledge management that combines both process and data driven approaches to ontology development and builds on them is described. In this methodology, called the BPD/D Ontological Engineering Methodology, competency questions that state the capability of an ontology to support business processes and decisions are specified. Concurrently, architectural requirements that specify aspects of existing systems that constrain ontology design choices are also stated, so that existing data repositories are explicitly considered and built upon when developing ontologies. Information systems tools to support both ontology-based knowledge management system construction and use can then be designed to support the steps of this methodology.