Adaptive Learning Systems Technology Advisor: A model for business entrepreneurs to implement IT

Software agents are the new paradigm shift in the quest to develop intelligent software. Their roots can be found in Artificial Intelligence (AI) and Human Computer Interaction (HCI), but differ from traditional AI techniques, which use symbolic knowledge representation to emulate "human-thinking" in computer systems. Agent-based systems are simply programs designed to simulate "intelligence" in software. The idea is to create personalized software. Although this type of software holds great promise for HCI and Information Technology (IT), this infant field lacks theoretical foundations and wide-scale acceptance practical for real-world problems. IT has been credited as the key to unlock benefits otherwise inaccessible to businesses. This compels a diverse population of business entrepreneurs to make IT implementation decisions. The process of IT adoption/use is critical to deriving its benefits. Yet, there are very few studies focused on the IT implementation pre-adoption phase. Software agents possess the characteristics that can assist decision-makers during this phase.; This dissertation studies a software agent, ALSTA (Adaptive Learning System Technology Advisor), in a theoretical framework to examine its effectiveness as a decision support system (DSS) during the IT adoption process. The model clearly defines ALSTA, proposes characteristics for ALSTA and describes a method for evaluation. It supports the works of earlier researchers who suggested that there is a need to include mediating entities as part of the technology acceptance model (TAM) [Agarwal et al., 1998, Gefan et al., 1998 and Karahanna et al., 1999].; Beyond the framework, this research empirically validates the effectiveness of ALSTA in a laboratory experiment. The experiment compares support rendered by ALSTA, a non-adaptive DSS (NADSS) and a human consultant (HDSS) during the IT implementation decision-making process. One hundred and ninety subjects participated in the experiment. The effectiveness of these systems was measured along several dimensions. The results strongly suggest that utilizing ALSTA during the IT implementation decision-making is worthwhile. ALSTA outperformed or performed comparatively to NADSS and HDSS on all measures observed. This role as a DSS that positively assists decision-makers should not be overlooked. These results are significant contributions to the theory/practice of software agents, HCI, IT implementation and DSS.