This paper presents a proposal for using data mining models to semi-automatically evaluate employee performance and address issues in areas such as talent attraction, selection and onboarding, performance evaluation, development and succession planning, training, job analysis and description, retention, and other responsibilities of the Human Resources department. The proposal is based on an automatic classification model, which achieves high accuracy compared to unsupervised learning algorithms. For the model development, an adaptation of the CRISP-DM methodology was used, allowing for the creation of a reliable model that can be integrated into cloud service platforms. The implementation of this model offers time savings in future evaluations of personnel in public organizations, is easy to apply, and provides result visualization for consultation and decision-making.