Queensland has 122 hospitals with over 1.8 million presentations to Emergency Departments (ED) each year. Each facility operates with different electronic medical records systems, making it challenging to identify which ED presentations are related to traumatic injury caused by traffic accidents. Identifying traffic accidents related injuries help with CTP (Compulsory Third Party) insurance claims. Manual identification of trauma cases typically took up to five hours per day, relying on keyword searches in ED presentation descriptions and diagnoses. This study aimed to evaluate the implementation of a new machine learning (ML) model to screen and identify ED admitted patients with traumatic injuries caused by traffic accidents. After building a unified data collection platform for ED presentations, manual identification was initially used as a reference point. The machine learning model was implemented using supervised classification and decision tree processes, such as: LightGBM and XGBoost. The traffic-related injury model achieved an accuracy of 98% classifying 167,541 ED trauma admissions. This artificial intelligence (AI) tool has redistributed screening hours, allowing clinical staff to focus more on quality improvement initiatives. The introduction of AI tools in trauma has proven to be valuable and clinical teams are now using the developed model, which has reduced screening manual task by 80–85%.