To support retail sustainability, retailers must also focus on energy efficiency. Cold storage is a major factor in the retail sector's rising energy use. Improving cold storage saves energy. Poorly filled or used cold storage wastes energy and increases costs. This situation pertains to retail inventory management. Therefore, retail inventory management is crucial. This study aims to develop an Artificial Intelligence (AI)-and IoT (Internet of Things)-based inventory management model to improve cold storage replenishment efficiency while considering energy costs. The suggested solution includes an inventory management model and application that alerts retail managers to quickly replace cold storage items and keep inventory levels steady. The suggested inventory management approach comprises two phases: development and execution. The development phase begins with data collection, using sensors and a microcontroller to monitor cold storage. The suggested framework incorporates the following data: 1) cold storage door opening and closing times, 2) electrical energy consumption, 3) cold storage temperature, 4) product weight, and 5) product photos. The Convolutional Neural Network (CNN) is applied to categorize visual input for generating refill alerts. Experimental results suggest that the proposed strategy effectively reduces total energy costs per unit by 18.1% and total inventory costs by 6.9% when compared to the typical periodic review inventory management model.