DESENVOLVIMENTO DE UM CHATBOT BASEADO EM RAG PARA APOIO AO ATENDIMENTO MULTIPROFISSIONAL NA EDUCAÇÃO PÚBLICA

The increasing complexity of social demands within the school environment has intensified the need for more structured institutional responses, particularly regarding the work of multidisciplinary teams in public education. In this context, persistent challenges are observed, including fragmented information, limited access to regulatory frameworks, and a lack of standardization in service provision, which undermines the effectiveness of actions and the reliability of institutional decision-making. In response to this problem, this study aims to develop a chatbot based on the Retrieval-Augmented Generation (RAG) approach to support multidisciplinary assistance through the systematization and retrieval of institutional protocols and normative documents. Methodologically, the research adopts the Design Science Research (DSR) approach, focused on the construction of a technological artifact capable of addressing real-world problems while integrating scientific rigor and practical relevance. The development process involved the organization of documentary databases, their structuring into computationally processable formats, and integration with language models, enabling semantic information retrieval and the generation of contextualized responses. The chatbot was designed as an informational mediator, capable of guiding professionals based on official guidelines, reducing reliance on individual interpretations and promoting greater consistency in procedures. The findings indicate that the use of RAG-based technologies contributes to the standardization of service workflows, enhances educational governance, and strengthens multidisciplinary practices by improving qualified access to information and supporting more reliable decision-making processes aligned with current regulations.