Sistema de apoio administrativo com agentes inteligentes para resolução de dúvidas sobre estágio

Abstract

This work aims to develop a system based on Artificial Intelligence Agents to assist the Internship Coordination in handling and resolving students’ inquiries related to supervised internship processes. To achieve this, a RAG system was implemented with the Qdrant vector database, enabling the storage and retrieval of documents such as regulations, laws, and guidelines. The agente architecture was designed to allow the registration of managers, agents, and tools, organizing and separating their specific functionalities. The solution was structured so that, upon receiving a user’s question, the agent accesses the database, retrieves relevant information, and provides a response through a chat system integrated with the Microsoft Bot Framework. Some limitations encountered included the dependence on the quality of the indexed documents and difficulties related to the automatic reformulation of questions. The results indicate that the system was able to provide clear and adequate answers with satisfactory performance, confirming the great potential for applying systems similar to the one proposed in this work.