Draping simulations are widely used for the development and optimization of manufacturing processes in the composite industry. These numeric simulations are typically performed using the finite element (FE) method, which are often quite time-consuming. To address these limitations, recent research focuses on surrogate modelling approaches based on artificial intelligence (AI). There are already various approaches implementing AI-based methods from simple approaches to sophisticated applications using neural networks and deep-learning based surrogates [1 – 4]. This study investigates the potential of surrogate modeling for approximating FE draping simulations. An FE model representing the draping of a glass fiber fabric onto a rib tool was developed using Abaqus/CAE and parametrized to enable automated training data generation. Details of model setup and data generation workflow are provided in [5]. Using this baseline FE model, two surrogate approaches of different complexity were analyzed and compared: an image-based surrogate model and an Extended position-based dynamics (XPBD) approach. For the image-based surrogate model, training data was generated by varying pressure magnitudes over three areas of the blank cut top surface. A U-Net was trained using these input parameters, with the FE results encoded as grayscale images. The second approach integrates an XPBD simulation with a reinforcement learning agent that optimizes draping path parameters to reduce wrinkle formation. To simplify the problem, the original blank cut was replaced by a square geometry. Simplified material properties were implemented in the XPBD model and calibrated against the FE baseline. Comparisons between FE and surrogate results showed that both approaches accurately predict the topology of the draped fabric. The U-Net offers fast predictions with minimal implementation effort and only slight deviations; however, it lacks geometry invariance. In contrast, the XPBD approach supports interchangeable geometries and offers greater generalizability, paving the way for the use of surrogates in composite manufacturing applications.