2026-04-302026-04-302025-12-18SILVA, André Oliveira Carvalho da. Chatbot inteligente para auxílio à tomada de decisão em empresa do setor de energia. Orientadora: Elen Priscila de Souza Lobato. 2025. 42 f. Trabalho de Curso (Bacharelado em Engenharia Elétrica e Biomédica) – Faculdade de Engenharia Elétrica e Biomédica, Instituto de Tecnologia, Universidade Federal do Pará, Belém, 2025. Disponível em: https://bdm.ufpa.br/handle/prefix/9503 . Acesso em:.https://bdm.ufpa.br/handle/prefix/9503This work presents the development of an intelligent chatbot designed to assist operational decision-making in an electric power utility company located in the Northern region of Brazil. The application employs Large Language Models (LLMs), semantic embeddings, and the Retrieval-Augmented Generation (RAG) technique to provide accurate, contextualized, and multimodal responses based on an indexed textual and visual database. The system architecture was structured using the LangGraph library, which enables dynamic control of the conversational flow, and Milvus, a vector data management system. The interface, developed with HTML, CSS, and JavaScript, provides an interactive and user-friendly experience. Tests demonstrated accuracy in the responses and an average response time of less than two seconds, highlighting the chatbot’s potential as a strategic tool for enhancing operational efficiency in the electrical sector. The study underscores the technical feasibility and positive impact of applying generative AI in critical industrial environments.Acesso AbertoSetor ElétricoRecuperação Aumentada por GeraçãoModelos de Linguagem de Larga EscalaTomada de Decisão OperacionalChatbotElectric Power SectorRetrieval-Augmented GenerationLarge Language ModelsOperational Decision-MakingCNPQ::ENGENHARIAS::ENGENHARIA ELETRICATrabalho de Curso - Graduação - Monografia