In today’s interconnected world, a wide range of devices, such as sensors, mobile platforms and smart appliances, continuously produce large volumes of data. When processed in real time, this data can reveal critical insights into system behaviour and environmental conditions. Complex Event Processing (CEP) systems are designed to extract such insights by identifying meaningful patterns in event streams. However, traditional CEP approaches rely heavily on domain experts to manually define detection rules, which limits scalability, introduces human error, and constrains the complexity of patterns that can be effectively captured. This paper proposes a novel approach that integrates Machine Learning techniques into the rule creation process for CEP systems. By automating the discovery of complex event patterns, our method reduces the dependency on expert knowledge and enhances the adaptability and reliability of CEP applications. We present the architecture of our system, demonstrate its application in a real-world scenario, and discuss the technological challenges involved in merging CEP with ML. Experimental results highlight the potential of our approach to improve pattern detection accuracy and system robustness in dynamic environments.