The challenge of Morocco is to provide low-cost housing and conserve natural sand. To overcome this challenge, it is necessary to find local alternatives. Moreover, the production of conventional concrete is a major source of CO2 emissions, which is a source of environmental degradation. In this paper, we studied a more eco-friendly concrete, consisting in a partial replacement of natural sand with shale from the Settat Basin. Different mixtures with replacement from 0% to 25% have been considered keeping a constant water-cement ratio 0.5. The methodology used is a combination of laboratory experiments and artificial intelligence techniques. Support Vector Machine (SVM), Random Forest (RF), and Deep Neural Network (DNN) models have been used to predict the mechanical performance of the concrete. Experimental results show that a replacement of 15% of the sand with shale is the optimum solution, which preserves 94% of the compressive strength (18 MPa) while greatly minimizing the thermal conductivity to 0.65 W/m.K. Among the prediction models, SVM stood out with the highest accuracy (R2 0.98). Additionally, the 15% shale mix worked well for 3D printing, showing stable extrusion and keeping over 90% of its mechanical strength after printing. Overall, our research shows that Settat shale can help create strong, eco-friendly concrete that works for both traditional building and modern 3D printing in Morocco.