Most methods used to solve partial differential equations require creating a mesh that represents the model's geometry. Today, unstructured mesh technology is widely used, allowing three-dimensional meshes with hundreds of millions of elements to be generated in just a few minutes. However, when optimising a design, many simulations are needed for different operating conditions and geometric configurations. Creating the best mesh for each setup becomes very time-consuming due to the requirement of excessive human intervention and expertise. This talk will cover our recent work on using artificial intelligence to predict near-optimal meshes suitable for simulations. The main idea is to take advantage of the large amount of data that already exists in the industry to improve the selection of a suitable spacing function, including anisotropic spacing. The proposed approach aims to use knowledge from previous simulations to guide the mesh generation process. I will assess the proposed method based on the accuracy of the predictions, efficiency, and environmental impact. This includes considering the carbon footprint and energy consumption of the computations required for a parametric CFD analysis under different flow conditions and angles of attack. References: [1] C. Lock, O. Hassan, R. Sevilla and J. Jones, Meshing using neural networks for improving the efficiency of computer modelling, Engineering with Computers, 39, 3791-3820, 2023. [2] C. Lock, O. Hassan, R. Sevilla and J. Jones, Predicting the Near-Optimal Mesh Spacing for a Simulation Using Machine Learning, Lecture Notes in Computational Science and Engineering, SIAM International Meshing Roundtable 2023, 115-136. Springer, 2024. [3] S. Sanchez-Gamero, O. Hassan and R. Sevilla, A machine learning approach to predict near-optimal meshes for turbulent compressible flow simulations, International Journal of Computational Fluid Dynamics, 2024.