Good point. I looked at the diffraction correction option, but was a bit put off when I learned it was a special form of sharpening being applied, when I don't really want to add sharpening till the very end of my processing pipeline. I could be wrong or misunderstanding this process option - most likely I am! - but that was my take away.
I agree most sharpening should be usually done at the end of the pipeline (at least when quality is a higher priority than workflow speed).
However deconvolution is not quite the same as a USM. It uses knowledge of the specific type of blur that occurs with a given lens-aperture combination (the point spread function) and attempts to restore the signal to (closer to) it's original state.
I can't tell you if it's right for your specific needs, or if another workflow would be better. But I do feel strongly it's worth looking at to make that determination yourself. In my
bodyscape work I've already increased the aperture I feel comfortable working with by a full stop as a result of Diffraction Correction, which sounds small, but every stop really helps!