In the world of professional football, the ongoing search for new talents is of paramount importance. Each club relies on scouts, who produce thousands of players’ reports every year. As a result, scout managers are faced with an overwhelming volume of data, primarily in the form of unstructured text. To explore and analyze such data, Large Language Models (LLMs) have recently emerged as promising tools, offering new possibilities for handling and interpreting these vast amounts of data. In this paper, we propose an LLM-based approach to assist scout managers in exploring players’ reports textual data. The approach leverages a Retrieval-Augmented Generation (RAG) architecture, wherein LLMs are employed for multiple tasks: (i) summarizing and extracting key information from players’ reports; (ii) interpreting scout managers’ input queries; (iii) generating the final response based on scout managers’ queries and on summarized players’ reports, the latter extracted from a vector database. The approach has been implemented within a service-oriented ecosystem, and validated in the scope of a real-world talent scouting use case, demonstrating its benefits in providing insights from unstructured players’ reports.