CHATBOT WITH ARTIFICIAL INTELLIGENCE FOR CONSULTING CBMPR FIRE AND PANIC SAFETY STANDARDS

Abstract

This article presents the development and validation of a chatbot prototype for consulting the Laws and Standards of the Paraná Fire Department (CBMPR). The prototype was configured to use hybrid search and Retrieval-Augmented Generation (RAG). The chatbot was able to answer several technical questions objectively, in a standardized and auditable manner, always correctly basing its answers on the applicable standards. Based on the positive results achieved in simulated tests, an architecture that allows for scalability and an implementation plan were proposed, which would enable its use by the general public. Expected outcomes include expanding and simplifying citizens’ access to public services, ensuring consistency in interpretation, and reducing search time, making the Chatbot a very useful tool.