This study investigated the use of Agentic Artificial Intelligence (AI) tools for evidence-based decision making among educational administrators in Cross River State, Nigeria. Specifically, it examined how educational administrators use Agentic AI for decision making, the influence of Agentic AI on the quality and effectiveness of administrative decision making, and the challenges and enabling factors affecting AI adoption. A descriptive survey design was adopted, and the population comprised 1,404 educational administrators, including principals, vice principals, directors, education officers, and administrative personnel. A proportionate stratified random sampling technique was used to select 311 respondents. Data were collected using a structured questionnaire and analysed using descriptive statistics (mean and standard deviation) and inferential statistics (correlation and regression). The findings revealed that educational administrators moderately use Agentic AI tools for evidence-based decision making, with the overall grand mean score of 3.31. The study also found that Agentic AI positively influences the quality and effectiveness of administrative decision making, particularly in improving accuracy and timeliness. However, AI’s impact on resource allocation and staff performance evaluation was limited. Furthermore, the study identified significant challenges to AI adoption, including poor ICT infrastructure, lack of training, and data quality issues, while leadership support was found to be moderate. The correlation analysis showed a strong positive relationship between AI use and evidence-based decision making (r = 0.62) and between AI use and decision quality (r = 0.58). Regression analysis indicated that infrastructure, skills, and organisational support significantly influence the effective use of AI, explaining 46% of the variance. The study concluded that although agentic AI has the potential to enhance educational administration, its adoption in Cross River State is constrained by infrastructural and capacity limitations. It recommended improved ICT infrastructure, training, data management, and leadership support to enhance AI adoption for evidence-based decision making.