<![CDATA[Conventional rural administration in Indonesia faces manual clerical bottlenecks, while existing digital solutions often fail due to the digital divide and an inability to process unstructured local dialects. To address these challenges, this study designed and evaluated an Intelligent Robotic Process Automation (IRPA) prototype for fully autonomous village correspondence, specifically handling seven distinct types of official letters. Utilizing the Design Science Research Methodology (DSRM), the proposed architecture integrates the WhatsApp Business API as an inclusive conversational interface, Google Gemini for cognitive intent classification, and the n8n low-code platform for cloud-based document orchestration. Functional evaluation using 20 test prompts, which represent real-world informal language, abbreviations, and typographical errors, demonstrated that the cognitive agent achieved 100% accuracy in intent recognition and boundary detection. Furthermore, the system significantly reduced administrative Turnaround Time (TAT) by approximately 77.5%, effectively transforming manual processes of 15 to 25 minutes into a 3 to 6 minutes automated cycle. Ultimately, this research offers three main contributions: an asynchronous architecture to lower user cognitive load, a deterministic prompt engineering method for public services, and empirical evidence of RPA efficiency in inclusive rural governance.]]>
Copyrights © 2026
