<![CDATA[Population administration services, particularly the issuance of electronic ID cards (e-KTP), still face the classic problem of insufficient clear information regarding requirements, procedures, and estimated completion times, often forcing the public to make repeated visits to the Population and Civil Registration Office (Disdukcapil). The City of Batam has implemented innovations such as the Electronic Population Service application (LAKSE), yet interactive information channels capable of answering citizens' questions directly, quickly, and contextually remain limited. The development of Large Language Models (LLMs) has opened significant opportunities for building public information service chatbots; however, generic LLMs are prone to hallucination—producing answers that sound convincing but are inaccurate or inconsistent with current regulations. This research aims to design and develop an LLM-based chatbot enhanced with a Retrieval-Augmented Generation (RAG) approach to accurately answer public questions related to e-KTP issuance in the City of Batam, since the generated answers are retrieved and referenced directly from an official knowledge base such as Disdukcapil regulations, administrative requirements, and LAKSE procedures. The development method employs a software engineering approach comprising the stages of document collection and cleaning, embedding into a vector database, integration of the retrieval mechanism with the LLM, and testing of answer accuracy using semantic similarity metrics and expert evaluation. The expected outcome of this research is a chatbot prototype capable of providing relevant, contextual, and accountable information, thereby reducing staff workload, shortening public waiting times, and supporting the transformation of digital public services in line with the implementation of the Electronic-Based Government System (SPBE) within local government environments.]]>
Copyrights © 2026
