Developing an AI-powered chatbot for enhanced user engagement in academic libraries

Thesis Abstract

In the contemporary digital age, academic libraries grapple with declining physical visits and the increasing demand for instant, personalized assistance from users accustomed to seamless digital interactions. Despite investments in digital catalogs and online services, user engagement remains suboptimal, partly due to limited responsiveness and lack of personalized communication channels. This study aims to develop and evaluate an AI-powered chatbot as a technological intervention to enhance user engagement in academic library settings. The specific objectives include designing a functional prototype of the chatbot, assessing its impact on user satisfaction and engagement levels, and identifying factors that influence user acceptance and interaction with AI-driven services. Employing a mixed-methods research design, the study integrates quantitative surveys and qualitative interviews to comprehensively explore user perceptions and behavioral responses. The population comprises undergraduate and postgraduate students registered at a major research university with a student body of approximately 40,000. A stratified random sampling technique selected a sample size of 600 students for quantitative data collection through structured questionnaires measuring engagement, satisfaction, and perceived ease of use. Complementary qualitative data were collected from 20 participants through semi-structured interviews to explore in-depth user experiences and attitudes towards the AI chatbot. Data collection instruments were pre-tested for validity and reliability, with Cronbach's alpha coefficients exceeding 0.85 across all scales. Quantitative data were analyzed through descriptive statistics, correlation analysis, and multiple regression to identify significant predictors of user engagement. Thematic analysis was employed to interpret qualitative interview transcripts, identifying common themes such as trust in AI, perceived usefulness, and interaction ease. The development phase involved creating a prototype based on open-source chatbot frameworks integrated with natural language processing capabilities and library-specific FAQs. The evaluation phase employed experimental design with pre- and post-interaction surveys to measure changes in engagement levels attributable to the chatbot. Analytical techniques included paired sample t-tests and ANOVA to assess the significance of observed differences. Anticipated findings suggest that the AI-powered chatbot significantly enhances user engagement and satisfaction by providing immediate, accurate responses and 24/7 accessibility. The regression analysis is expected to reveal perceived ease of use and trust as critical determinants of user acceptance, while thematic analysis is likely to highlight dimensions such as usability, trustworthiness, and personalization as central to user interaction. The study is expected to demonstrate that integrating AI-driven assistance in academic libraries can address existing service gaps, foster increased user interaction, and contribute to improved information retrieval experiences. This research contributes to the body of knowledge by providing empirical evidence on the efficacy of AI chatbots in fostering user engagement within academic library contexts, guided by the Technology Acceptance Model (TAM) and the Unified Theory of Acceptance and Use of Technology (UTAUT). It offers a practical framework for designing, implementing, and evaluating AI-driven library services that are aligned with user expectations and needs. In conclusion, the study recommends the adoption of AI chatbots as strategic tools to transform traditional library services, emphasizing ongoing user training, system enhancement based on user feedback, and further research into long-term impacts on library engagement metrics. The findings hold implications for academic institutions seeking to modernize library services and adopt innovative ICT solutions to meet evolving user demands.