Abstract

In this study, we explore the integration of Artificial Intelligence (AI) in responding to questions related to Islam. We hypothesized that utilizing a Large Language Model (LLM) alongside a comprehensive knowledge base and human oversight could streamline the question-answering process. We designed a system architecture and implemented it in a case study project within this specific domain. To assess the efficacy of our system, we processed 140 real-user queries. The findings suggest that an LLM, when supplemented with a verified knowledge base, significantly reduces the workload on human experts by generating preliminary responses for further refinement by domain experts.