The new feature uses a pipeline of machine learning models to instantly analyze equipment behavior and identify triggered warnings and anomalies within a selected time interval. Instead of relying on a single model, the system combines the results of several specialized models to provide a more reliable assessment of the detected event.

In most cases, the analysis pipeline consists of three to four machine learning models, each providing its own confidence score for the detected condition. By combining these independent predictions, the system can determine whether the models agree on the presence of an anomaly. When all models identify the same abnormal pattern, the event can be confirmed with up to 95% confidence.

This approach allows engineers to analyze large amounts of historical and real-time data much faster, automatically filtering out insignificant deviations and highlighting events that require attention. By combining multiple AI models into a single analysis pipeline, nv.ebs turns complex equipment data into clear and actionable insights within seconds.