From a multiagent viewpoint, a workflow is a dynamic set of tasks performed by a set of agents to reach a shared goal. We show herein that commitments among agents can be used to model a workflow and coordinate their execution of it. From a service-oriented computing viewpoint, a workflow can be represented as a set of services and a specification for the control and data flows among these services to address some business needs. As a formal declarative knowledge representation model, ontology is used as a basis for agent-based workflow execution and coordination. This paper presents methodologies to map an OntologyWeb Language for Services (OWL-S) representation for a workflow to a CPN graph, a graphical and mathematical modeling tool for describing and analyzing information processing systems, and then infer commitments and causal relationships from the CPN graph. We provide an example scenario to describe our algorithms.