A module was developed to assess investment risks in virtual IT companies using machine learning, reducing time spent on risk evaluation for investors. The algorithm was based on expert evaluation of 23 risk parameters from 20 projects. Eight machine learning methods were used to predict investment risks, with Support Vector Classifier, Random Forest Classifier, and K-Neighbors Classifier showing the best forecasting performance. The study aimed to support decision-making in virtual IT companies by identifying and evaluating key risk factors.