This paper evaluates the use of machine learning models for text classification as a means to reduce the time required to construct the Coverage Indicator in the Technical Products Department of the Superintendency of Electricity and Fuels (SEC) in Chile. The current manual process to identify whether imported electrical products are certified is time consuming and limits national inspection capacity. The proposed solution uses supervised machine learning models implemented in Python, including Support Vector Machines, Random Forest, Logistic Regression, and others, to automatically classify textual descriptions of imported products. The results of a pilot study demonstrate that the classification time can be reduced by more than 90%, with an accuracy rate that exceeds 99% using SVM models. In addition, the application of these models led to a significant improvement in the coverage indicator and enabled a wider coverage of inspection at the national level. The study confirms the hypothesis that artificial intelligence techniques can substantially optimize regulatory processes, reduce operational times, and enhance public safety through faster detection of non-compliant products.