

Semantic Enrichment: Through the use of LLMs, descriptive metadata was generated to provide a complete and multidimensional understanding of datasets. This enabled the creation of a semantic representation suitable for integration into a vector database.
Virtual Assistant: The chatbot orchestration is managed through direct integration with the ISTAT website, including workflows for request routing and autonomous handling of more complex tasks such as graphical data visualization.
Benefits: The solution simplified access to datasets and improved data navigation both in terms of response quality and website usability, while also strengthening ISTAT’s SEO positioning.