<![CDATA[Asset management in information technology at the IT Division of PT Pertamina Hulu Rokan currently faces scalability challenges due to reliance on conventional manual spreadsheet-based processes in managing approximately 30,000 asset data. This condition triggers the risk of human error, data redundancy, and time inefficiency in preparing asset condition recapitulation reports. This study aims to design a website-based Data Summary System that integrates Large Language Models (LLM) technology and spreadsheet data through the use of the Model Context Protocol (MCP) as a form of secure and scalable integration standard. System development was also carried out using a two-iteration prototype method, with Microsoft Azure infrastructure, the FastAPI framework (Python) on the backend side, and React.js on the frontend side. System quality testing was carried out based on the ISO/IEC 25010 standard which covers the eight main characteristics. The test results showed 100% achievement in the functional suitability aspect and the highest rating (Rating A) in the reliability, maintainability, and security aspects using SonarQube analysis. In terms of performance, the system demonstrated high responsiveness with a GTMetrix score of 100% and a usability level of 80.9 (Good category) on the System Usability Scale (SUS). The effectiveness of the system was tested through a One-Group Pretest Posttest experimental design on 8 IT employee respondents. The analysis results showed a good increase in productivity with a Normalized Gain value of 0.91 (High category) and an effectiveness percentage reaching 91.48% (Effective category). Therefore, this Data Summary System has proven successful in minimizing repetitive work steps and increasing accuracy and speed in managing IT assets at PT Pertamina Hulu Rokan.]]>
