This study explores the potential of generative artificial intelligence (GenAI) in enhancing the university recruitment process by integrating AI-driven chatbots to assist applicants. Focused on Krakow University of Economics, the research evaluates five language models for their effectiveness in providing accurate responses to admission queries. Utilizing a blend of lexical, semantic, and model-as-judge evaluation methods, the study demonstrates varied performance across models, with significant implications for enhancing candidate experience and streamlining communication. The findings highlight the need for iterative improvements and domain expert interventions in refining AI responses, aiming to optimize the integration of GenAI in higher education settings. The research contributes to the understanding of AI applications in academia and suggests practical approaches to improve information dissemination during critical recruitment phases.