Government institutions are facing a significant challenge in handling public inquiries effectively. This often leads to delays and inefficiencies in delivered service. The problem is exacerbated by the increasing number of inquiries received through government websites. The complexity and vulnerability of human classification processes to errors are evident. This work proposes a machine learning model to automatically assign customer requests to predefined categories based on their characteristics. The proposed model aims to automatically classify customer requests and accurately categorize inquiries into appropriate departments based on various criteria, including inquiry type, date, and the person inquiring. This will help to reduce the errors caused by humans, reduce delays, and increase the efficiency of service delivery. In this paper, different machine learning models have been developed, including XGBoost, Support Vector Machine, Linear Support Vector Machine, Logistic Regression, Gradient Boosting, Decision Trees, Random Forest, and K-Nearest Neighbors coupled with Term Frequency-Inverse Document Frequency. Extensive computational experiments were conducted on the Arabic dataset collected from the Directorate of Housing Sharjah (DH Sharjah) with approximately 13,810 records. The results show that the Support Vector Machine achieved the best result among all other models, with an accuracy of 79% using five-fold cross-validation.