Chatbot para responder a las preguntas frecuentes sobre el proceso de nivelación y admisión de la Universidad Técnica de Machala

Abstract

Abstract: In this work, a conversational chatbot is developed for the Technical University of Machala, which provides automated responses to questions about leveling and admission processes, mitigating the saturation of traditional service channels and the lack of availability of administrative staff. This solution eliminates barriers of schedules and geographic location by implementing itself on a hybrid architecture that integrates various tools such as Rasa for dialogue management, ChromaDB for semantic retrieval, Mistral for generating natural responses, and PostgreSQL for storing conversation history. To validate the experience it offers to users, the System Usability Scale (SUS) questionnaire is used, and its score of 79.3/100 indicates an acceptable level of usability with a tendency to be a promoter. The technical efficiency along with these results demonstrate that it is highly viable to create inhouse solutions in public universities using open-source tools, which guarantees technological sovereignty, control over institutional data, and cost reduction. Additionally, the implemented architecture ensures adaptation to growth and represents a model that can be replicated in other educational institutions.