This chapter presents an agent-based architecture for integrating organizational knowledge repositories and business processes orchestrated by a workflow management system. This architecture proactively provides relevant knowledge to workflow tasks considering their context, and stores the information generated by its execution for future requirements. It describes components of the architecture, models a multi-agent system that enables the integration, presents a strategy to annotate and retrieval knowledge of non-structured information sources, and defines a new workflow pattern to be used in knowledge intensive tasks in order to make possible the knowledge provision. This architecture allows workers to count, in a proactive way, with all necessary information for the task executions without suspending their activities to retrieve information scattered in the organization. It reduces the wasted time in manual knowledge searches included in mostly knowledge management approaches.
