The need to reduce time and costs in the construction industry, specifically in the field of concrete, has turned its attention to non-destructive testing as an opportunity to estimate the compressive strength of concrete. However, the interpretation models for these results have shown some issues with reliability and accuracy. This is why this research proposes developing a compressive strength estimation model using artificial intelligence tools like neural networks and maximum sensitivity. This is for concretes of 20, 25, 30, and 40 MPa at ages of 3, 7, 14, 28, 56, and 90 days. For these, rebound hammer, ultrasonic pulse velocity, and the corresponding response, compressive strength, were measured. With these results, a linear regression reference model, an artificial neural network model, and a maximum sensitivity model were developed to reduce the error between actual and predicted values. The results help us form better expectations of AI tools and consider using them as complexity increases (more variables, more responses).