This study addresses the challenge of accurately monitoring and assessing energy usage in commercial buildings. In many buildings, anomaly detection due to faulty conditions of sensors, actuators, or controllers is reactive and requires significant time and effort. The proposed solution addresses the problem using pre-existing information, without necessitating additional resources. The methodology involves collecting and processing electricity demand data, segmenting it into weekly intervals, and identifying typical usage patterns through clustering. Representative clusters define a building’s standard weekly demand profile, serving as a benchmark for monitoring and comparing demand patterns. To enhance the monitoring process, threshold-based alerting mechanisms ensure timely identification of significant deviations in usage patterns. If a significant deviation is detected, it is compared to the representative profiles of similar buildings. If these buildings exhibit the same deviation, it may indicate drastic changes in weather conditions, seasonality, or extremely high pricing, causing all customers to alter their behavioral patterns. In contrast, if the patterns of similar commercial buildings remain unchanged, this explicitly indicates problems within the specific building. Ultimately, our approach not only enhances the precision of demand monitoring, but also facilitates early anomaly detection, optimizing energy usage, and reducing operational costs to enhance the sustainability and efficiency of commercial buildings.