The digital twin approach aims to improve decision-making by providing more accurate predictions of the behavior of the modeled asset. Even though the concept has gained popularity in the Architecture, Engineering, Construction, Operations, and Management (AECOM) industries as a solution to existing challenges, its application remains limited. One reason is that traditional digital twins can only integrate deterministic input parameters and models. This restricts prediction accuracy, especially for geotechnical design and assessment with AECOM, where uncertainties are particularly large. For example, physics-based models are commonly used to predict soil settlement under load. They take soil parameters describing the mechanical properties of the soil and expected loads as input. However, geotechnical engineering data is often sparse and low-quality. Additional uncertainty is introduced when interpolation models are used to estimate values in locations where data is unavailable. However, these large uncertainties are not explicitly considered in most geotechnical projects. Design decisions are made for worst-case deterministic models to ensure safe design. Uncertainties are only considered indirectly by adding safety factors afterward. This approach can lead to overly conservative designs and inefficient resource usage. To address this challenge, we propose a Probabilistic Digital Twin (PDT) framework, which extends traditional digital twin methodologies by integrating uncertainties and is tailored to the requirements of geotechnical design and assessment. The PDT framework provides a structured approach to capturing various sources of uncertainty, including aleatoric, data, model, and prediction uncertainties, and propagating them throughout the modeling process. Bayesian methods allow the PDT to update models in real-time as new, site-specific information is obtained. The framework then facilitates a formal analysis and optimization of the sequential decisions that are taken under uncertainty in the construction process. The effectiveness of this probabilistic framework is demonstrated on a highway foundation construction project [1], showcasing its potential to improve decision-making and project outcomes amid significant uncertainties. REFERENCES [1] Bismut, E., Cotoarbă, D., Spross, J. and Straub, D., 2023. Optimal adaptive decision rules in geotechnical construction considering uncertainty. Géotechnique, pp.1-12.