Programa del congreso

This paper presents the design, implementation, and evaluation of a humanistic administrative chatbot developed for the Universidad de Ciencias y Humanidades (UCH) in Lima, Peru. While most AI-driven administrative systems in higher education prioritize efficiency, scalability, and response accuracy, this study proposes a complementary paradigm centered on relational accompaniment, ethical mediation, and institutional identity. Grounded in contemporary discussions on AI governance, large language models, and human-centered education, the chatbot was configured using Chatbase with a structured prompt engineering strategy aligned with UCH’ s educational principles. A quasi- experimental evaluation was conducted across two academic semesters (2024-2 and 2025-1), involving 85 students interacting with simulated administrative queries. Results indicate improvements in response time and user satisfaction, alongside high perceived warmth, clarity, and institutional alignment. The findings suggest that administrative AI systems can transcend purely functional roles and become instruments of formative mediation when designed under explicit humanistic principles. This model contributes to the discussion on responsible AI deployment in higher education, particularly in institutions committed to integral formation and social responsibility.