GoSAFE RAIL project is providing a means of virtually eradicating sudden failure on railway transport network by moving from considering critical infrastructure as inert objects to being intelligent (self-learning objects). The Safety framework for rail sector developed in the project is allowing data on transport network performance (e.g. visual inspection data, historic failure information, traffic data, databases etc.) to be stored in a Building Information Modelling, BIM environment. This information is fed into a decision support tool (DST) that enables decisions to be made in real-time allowing emergency safety measures and maintenance scheduling to be prioritised for infrastructure elements exhibiting stress. Algorithms at an object level (e.g. a level crossing, bridge/earthwork or tunnel) and at a network level (traffic flow model and DST) are incorporating machine learning to train the system to evolve with time using available monitoring data. Historical and forecast climate data are used as input to the system. The precision of monitoring and predictions are improving with time as the model predictions are compared to the object and network performance.