I’m confused. When I do post projects I typically export so that I get exactly the channel configuration I want. The way I do it is I route any desired (sub)mix of audio to an output which has no physical output attached to it, and then I select those outputs for my mixdowns.
So I will in some cases for example run all my dialog into a stereo group which is the last stage of processing, and then using the direct outputs I will send that dialog both to the main stereo mix, the mix minus narration which is also stereo, but also to the “dialog mono” output. That way I get exactly what I need.
Are you looking for something else?