Train a model for upscaling and enhancing normal maps

I use Photo AI a lot for upscaling textures for video game development. Because of the more precise detail needed for normal maps it would most likely require specific training or a dedicated model. It would be invaluable if this could be implemented in the future. If implemented, you could probably do a targeted advertisement campaign to game developers selling Photo AI as the new standard in game texture upscaling and enhancement $$. Thank you for any consideration!