The rise in online shopping has made it much easier for criminals to commit fraud, so new ways to find and stop it are needed. This paper presents a new way to improve fraud detection in retail settings by combining cloudless AI solutions with AIOps (Artificial Intelligence for IT Operations) frameworks. Our method uses edge computing architectures to process transaction data on-site. This maintains a high level of detection accuracy while lowering latency and increasing privacy. The proposed system employs adaptive learning mechanisms, real-time anomaly detection, and machine learning algorithms to uncover fraud tendencies independently of cloud infrastructure [8]. We tested our strategy using a dataset consisting of 2.3 million retail transactions. It is 97.4% accurate and only gives a false positive 0.8% of the time. This one is 23% better than the usual cloud-based options. Using cloudless architecture, the average reaction time went down from 850 ms to 120 ms, all while still following data sovereignty regulations. This study contributes to the expanding body of research on how AI may be used on the edges of networks. It also allows organizations to improve their fraud prevention without sacrificing efficiency or compliance with the law.