In recent years, local governments have increasingly sought digital solutions to improve the accessibility and efficiency of public services. However, many sub-district offices in Indonesia still rely heavily on manual service processes, which limit access to administrative information outside of office hours. This study aims to design and implement a web-based chatbot system to address information service challenges at the sub-district level. Using a descriptive-engineering approach, the system was developed following the waterfall model and evaluated through simulated user interactions. The chatbot was tested with 20 frequently asked questions related to key administrative services, including ID card (KTP), family card (KK), and domicile letters. The results show that the chatbot successfully answered 16 out of 20 questions (80%) with an average response time of 1.7 seconds. Most failures occurred due to variations in question phrasing not recognized by the system. This research demonstrates that chatbots can enhance basic information delivery in low-resource government settings. The findings contribute to the growing body of knowledge on digital public service transformation and provide a replicable model for implementing low-cost, scalable chatbot solutions in other local administrative contexts.
