<p>As an industry, the development of Artificial Intelligence (AI) is also accelerating the human resource (HR) management by allowing the use of data-driven information about employee engagement and work performance. The existing research, though, is mostly based on surveys or structured measures and therefore provides little insight on the latent conceptual relationships that determine the research of AI-enabled HR. This paper fills this gap by applying a new analytical framework based on natural language processing (NLP), Word2Vec technique. Word2Vec using KNIME software was used to analyze a dataset of 405 peer-reviewed articles extracted from Scopus database between the year 1989 and 2025. In the proposed solution, the textual data is converted into word embeddings of high dimensions and the use of K-means clustering is used to reveal the latent semantic relationships between employee engagement and work performance in AI-based HRM literature. The results indicate the different semantic clusters that depict the significant HR dimensions of personalization, wellbeing, productive, automation, and ethical governance. In contrast to the basic frequency-based or sentiment based models, this paper provides an example of how word embeddings over AI can reflect conceptual distances and novel research trends. The study makes a contribution to the field of semantic analysis by presenting a scalable framework of semantic analysis and a contribution to the practical sphere of ideas and information which can contribute to evidence-based HR decision-making in AI-based organizations.</p>