A Multi-Agent Knowledge Management System for Reactive and Proactive Knowledge Supply

Introduction

Knowledge is currently recognized as a valuable asset for organizations. It is considered a capital that has an economic value for an organization; a strategic resource used to increase productivity and offer stability in dynamic competitive environments (Ermine, 2000). In this new paradigm, in which physical capital and work are no longer the only fundamental basis for successful management, Knowledge Management (KM) has captured the attention of organizations as one of the most promising ways to reach success in this information era (Malone, 2003). Companies are beginning to understand the importance of knowledge in the organization as a resource which enables them to gain a sustainable competitive advantage (Bolloju et al.,2002).

In the pursuit of strategic competitiveness, great KM effort has dealt with capturing, codifying, storing, and disseminating knowledge in the organization (Lai, 2007). However, most of these efforts have been carried out in an isolated manner, without linking the organizational knowledge with organizational business processes. Nowadays, organizations coordinate their activities through business processes. To accomplish the tasks involved in a business process, workers can need information that is scattered along the organization.

This need for information should be considered in a contextualized way and be suitably delivered by the organizational KM infrastructure. Business Process Management (BPM) and KM are closely interrelated and are both essential elements of an organization. Therefore, this interrelation must be addressed in the development of a system architecture for KM.

Traditionally, KM approaches focused on providing information in a reactive way (Ale, 2009). When a worker needs information while executing a task, the worker must do an explicit search in an organizational knowledge repository with the aim of finding relevant information to continue the task execution. Once a task has been executed, the information generated as a result of its execution should be stored in a suitable way for future requirements. This paradigm is adequate in organizational areas in which the executed processes are unstructured, dynamic, unpredictable, and constantly changing without a fixed model that guides the control flow among their activities and workers. However, in areas where workers execute tasks belonging to a structured business process, information needs should be satisfied in a proactive way, allowing workers continue task execution while collecting all the necessary information from the organizational knowledge repository.

Workers will not do extra work to support a KM system unless it contributes to improve their jobs. In this case, organizational activities structured through a business process and automated by a Workflow Management System (WfMS) (Hollingsworth, 2005) should be considered as an opportunity to provide a KM infrastructure in order to bring the right knowledge to the right people in the right form and at the right time. WfMS should be considered a trigger of KM-support activities and a distributor of organizational knowledge that provides workers with necessary information to make better judgments and decisions.

In this new vision, process-based KM emerges as a potential challenge to support business processes with manual or semi-automatic tasks that need knowledge for their execution and, as a result, can produce new knowledge. Tasks with these needs are called Knowledge Intensive Tasks (KIT) and workflows composed of KITs are called knowledge intensive workflows (Sarnikar & Deokar, 2010). Process-based KM takes advantage of the task context provided by the workflow to deliver task-relevant knowledge. Organizations often execute activities with both types of information needs (reactive and proactive); thus, the KM architecture should support both needs (Jung et al., 2007).

In this chapter, we propose an agent-based architecture for an integration of business processes orchestrated by a WfMS with the organizational knowledge repository administered by ontology-driven KM systems. This architecture proactively provides relevant knowledge to tasks, considering their context and stores relevant knowledge in a proper way for future requirements. In case of necessity of more information, also the architecture reactively brings knowledge through natural language queries. Conventional workflows are extended with KM-support automatic tasks that are responsible for retrieving knowledge and storing results.