Intelligent-Agent-Supported Enterprise Information Portal

Abstract

Enterprises in today’s highly competitive environment must cope with mountainous information. At the same time, they must act quickly to make timely business decisions. Providing the right information at the right time has been a challenge for information technology professionals. Wells, Sheina, and Harris-Jones (2000) found that in the average company in the United States, 80% of information is kept on individual personal computers, and less than five percent of employee knowledge is captured and made available in enterprise systems. Various systems, such as management information systems, enterprise resource planning, and knowledge management systems have been used to support the information needs of organizations. In recent years, enterprise information portal (EIP) have been developed to provides an effective platform for knowledge workers in organizations to find, manage, use, and share previous disparate information. An EIP gives the user a single point of access to data, information, and knowledge across the entire enterprise, thus holding the promise of increasing the competitiveness of the organization (Mitchell, 2005; Schroeder, 2000). Aneja et al. (2000) proposed a generic framework for an enterprise portal that aimed at taming content chaos on corporate intranets. Raol, Koong, Liu, and Yu (2002) developed a simplified model that is used for classifying enterprise portal features and functions. In most EIPs, the key functions and features include interface, content management, search, collaboration, personalization, security, network, administrative tools, and extensibility. Typically, EIPs offer a familiar and easy-to-use Web interface to the users. However, a traditional Web interface lacks the breadth and flexibility desired by end users. Artificial intelligent agents have recently become capable enough to be integrated with enterprise information portals. Those agents can be used to facilitate the effective use of many of the features and functions in EIPs. Various types of intelligent agents, such as interface agents, personal assistant agents, information foraging and filtering agents, collaborative filtering agents, decision support agents, and bargaining and negotiation agents can work separately or collectively to make EIPs more accessible and user friendly. However, there are many theoretic as well as practical issues that must be addressed before intelligent agents should be widely deployed. We must understand the inherent limitations of intelligent agents and be able to assess the benefits and risks associated with using intelligent agents in EIPs. A report published by KPMG in March, 2000, found that 70% of the firms surveyed indicated that they experienced information overload even when enterprise information portals/intranets were used. Apparently, making vast information available is not enough in creating effective enterprise portals. We need to provide a set of intelligent functions to improve the effectiveness of EIPs. Traditional EIP functions from major EIP software packages are discussed in Raol et al. (2002). We attempt to bridge the research in enterprise portals and intelligent agents by identifying and characterize the role of intelligent agents may play in EIPs. Previous research has shown that intelligent agents can be used in various applications that have similar characteristics to EIP functions. For instance, intelligent agents are used in locating expertise in computer networks (Vivacqua, 1999) and in risk monitoring (Wang, Mylopoulos, & Liao, 2002). As intelligent agent technology becomes more mature, the application of intelligent agents has become more wide spread (Kontolemakis et al., 2004; Xu & Wang, 2005). However, there is little research on systematic integration of intelligent agents and enterprise information portals.