AI adoption readiness among Ukrainian education managers: Barriers, typologies, and policy implication

Artificial intelligence (AI) adoption in education requires strategic leadership from education managers, yet little research examines AI readiness among this stakeholder group, particularly in post-Soviet transitional contexts. As decision-makers who control whether and how AI enters educational practice, managers’ readiness is essential for evidence-based policy. This study investigates (1) the current state of AI adoption readiness among Ukrainian education managers, including the gap between personal and institutional readiness; (2) the barriers and facilitators that most strongly influence AI readiness; (3) distinct manager typologies based on readiness profiles; and (4) policy implications for scaling AI in education. A cross-sectional survey was administered to 395 education managers across 23 Ukrainian regions in September 2025. The 47-variable instrument covered demographics, AI readiness (personal and system), adoption barriers, tool usage, application potential, and psychological attitudes. Analysis employed descriptive statistics, Latent Class Analysis (LCA) for person-centered typology identification, fuzzy-set Qualitative Comparative Analysis (fsQCA) for configurational analysis of barrier combinations, and exploratory structural analysis. Personal readiness ( , ) significantly exceeded system readiness ( , ), a 0.68-point personal–institutional gap confirmed by a Wilcoxon signed-rank test ( , ). Regulatory absence and digital competency gaps were the most prevalent barriers. ChatGPT was the most widely adopted tool. LCA identified six provisional manager profiles, with the Competency-constrained class being the largest. fsQCA found that no barrier was a necessary condition and no barrier configuration was sufficient for low AI readiness, indicating that low readiness is not reducible to specific barriers or their combinations. Digital competency showed the largest training gap. Ukrainian education managers demonstrate higher personal than institutional AI readiness, constrained primarily by regulatory uncertainty and competency deficits. The identification of six provisional typologies enables targeted policy interventions. Policy priorities should focus on regulatory framework development, institutional capacity building, and differentiated, typology-based training.