Transformação digital no setor automotivo: desenvolvimento de um chatbot com modelos de linguagem para a extração de conhecimento de manuais automotivos

Abstract

With Artificial Intelligence (AI) continuously redefining the technological paradigm, its application in the automotive sector has become increasingly impactful, particularly in optimizing self-diagnostic and vehicular navigation systems. Thus, in this context of innovation, this study aims to simplify access to and interpretation of vehicle manuals through a chatbot powered by Large Language Models (LLMs). For this purpose, the adopted methodology encompasses everything from technology selection to chatbot implementation. In this process, the Milvus vector management system was used for data indexing and searching, while Streamlit was employed to create an interactive interface. Additionally, the solution architecture was containerized using Docker, ensuring ease of deployment and portability. Finally, the achieved results demonstrate that the chatbot can interpret natural language queries and provide accurate and relevant information from the manuals, thus contributing to transforming complex texts into simple interactions.