Protecting confidential information in energy companies is a critical task with serious national security implications. The article emphasises that inadequate protection can lead to serious consequences, including damage to national infrastructure and even military conflicts. The article analyses three levels of information protection: legal, organisational and technical. Particular attention is paid to the growing threat of insider intruders, which has become increasingly important in recent years. The research is aimed at using mathematical methods to estimate the probability of an attack on confidential information by an insider to identify priority areas for prevention and threat reduction. A multivariate regression-based model is proposed that takes into account many factors that influence the probability of an attack, such as the level of access to information, motivation, knowledge and skills of the intruder, as well as the probability of detection and punishment. To predict the probability of attack, we propose to use the logistic regression method, which allows predicting values in the range from 0 to 1. The model takes into account the interaction between the parameters and their mutual influence on each other, helping to develop a comprehensive approach to protecting confidential information and reduce the risk of information leakage in energy companies.