Cryptocurrency exchanges have become crucial platforms for trading and investing in the fast-changing field of digital finance. Nevertheless, due to the significant increase in their popularity, these exchanges have become highly profitable targets for fraudulent activity. This study presents “Scorecard Surveillance,” a novel methodology aimed at identifying and mitigating fraudulent activities within bitcoin exchanges. Our solution utilizes unsupervised machine learning techniques to systematically examine transaction patterns and allocate risk ratings according to abnormal behaviors. The scoring system is calibrated with previous data and expert insights, providing a dynamic, scalable, and effective method to promptly detect suspicious activities. This study showcases the effectiveness of the Scorecard Surveillance methodology in effectively identifying fraudulent transactions while reducing the occurrence of false positives. This research not only provides a strong tool to improve the security of cryptocurrency exchanges, but also establishes a basis for future investigations in the field of detecting fraud in digital money.