This paper presents a comprehensive computational analysis of 25 years of research published in the proceedings of the Mexican International Conference on Artificial Intelligence (MICAI), the most influential AI conference in Mexico with a growing international reach. Using a combination of bibliometric techniques and unsupervised natural language processing methods such as topic modeling, the study traces the evolution of research priorities, collaborative networks, and the tone of scientific discourse over time. The analysis identifies key topics and shifting patterns in the focus of AI research, from early work in logic and symbolic reasoning to more recent trends in machine learning, applied AI, and generative AI supported by large language models (LLMs). In addition to its retrospective contribution, the study underscores the importance of systematic, meta-level analyses in understanding and shaping national scientific agendas. By offering a data-driven portrait of MICAI’s intellectual trajectory, the work serves as both a historical record and a strategic resource for the Mexican AI community. It provides insight for researchers, policymakers, and institutions seeking to strengthen Mexico’s position in the global AI landscape. The framework presented may also support the study of scientific evolution in other national and regional AI communities.