This paper presents the findings of a case study exploring the HVAC flexibility potential of a mid-size office building in Prague. A cloud-based supervisory control algorithm was developed to communicate with the building management system (BMS) and was deployed for a three-month testing period. During this time, data from the day-ahead electricity market were used as a virtual price signal to optimize HVAC operation through load shifting strategies. A model predictive control (MPC) algorithm adjusted chiller operation every 10 min over an 8-h prediction horizon, aiming to reduce operational costs while maintaining indoor environmental quality. Following the testing period, long-term operational data were analyzed, revealing promising results. The findings demonstrate substantial energy and cost savings, with the chiller’s electricity consumption decreasing by over 25% and energy costs dropping by nearly 30% compared to the reference year. These outcomes highlight the potential of predictive control strategies to shift energy use to lower-cost periods while ensuring indoor comfort. The study validates the effectiveness of smart HVAC control in real-world applications and provides a foundation for scaling similar strategies to other commercial buildings.