Decision making in geotechnical engineering is characterized by considerable uncertainty. To find optimal solutions in such an environment, an iterative decision-making process, which includes new information as it becomes available, is required. In this contribution we extend the risk-based framework of Bismut et al. (2023) for optimal planning of a geotechnical construction to include predictions, which are based on a linear regression analysis of monitoring data. This approach is illustrated for the design of a surcharge on an embankment in clayey soil, as the optimal preloading sequence is searched. We demonstrate that by increasing the amount of information considered more cost efficient strategies can be identified and outcome uncertainties can be reduced.