It would use a sampler and for the subtle echo effect add a reverb to the bus.

https://www.youtube.com/watch?v=EQdp2QLiSYQ&t=187s

You could have AI do some postprocessing. I think a similaar approach is the future for image generation, you have a model output a 3D scene, use a classical raytracer to do rendering and then have a final model apply corrections to achieve photorealism.

the AI designs and controls the effects chain and mastering too