In the realm of economic policy, the communications from key financial institutions like Sveriges Riksbank significantly influence market behaviors and expectations. However, the extensive minutes from the Riksbank meetings, dating back to 1999, present a challenge for extracting specific insights due to their sheer volume and detail. This thesis addresses this challenge by developing a specialized system that leverages semantic chunking and natural language processing (NLP) techniques to streamline the analysis of speech data from the Riksbank representatives. The system organizes data into coherent segments, enhancing the retrieval, searchability, and usability of institutional textual data. The tool's design integrates principles of user experience (UX) and visual usability, validated through user testing and the System Usability Scale (SUS). This approach aims to provide a precise and contextually accurate tool for economic policy analysis, thereby improving decision-making processes. The results indicate that the tool is user-friendly and effective, though further enhancements are suggested for broader applicability and improved functionality.