Anomaly detection is a fundamental task in various domains, such as cybersecurity, finance, healthcare, and sensor networks. Traditional methods often struggle to distinguish between normal and anomalous behaviors when contextual information is not properly considered. This paper explores context discovery as a key strategy for enhancing anomaly detection. By identifying and utilizing relevant contextual information, anomaly detection systems can more effectively differentiate between benign and anomalous patterns, improving both the accuracy and robustness of detection. We present an approach to context discovery, where contextual variables such as time, location, or user behavior are dynamically extracted from the data, and how they can be incorporated into existing anomaly detection algorithms. We demonstrate the effectiveness of our method through a series of experiments on synthetic and real-world datasets, highlighting improvements in detecting anomalies in complex, context-dependent environments.