I record my podcast on riverside.fm and it outputs each person on the podcast as a video with only that person's video and audio. I upload these files into Descript then transcribe them as a single multi-track sequence. This both saves on transcription cost (it's only one file) and makes it easy to edit later in the process.
Every episode project has a composition titled "Episode N" with the multi-track sequence of the podcast participants. This is where all of the main editing for the podcast happens.
While the main composition is great for doing all of our work in, it can become hard to see whats gone after a lot of editing. To accommodate for this I create a "Raw Video" folder in the project. This folder houses an unedited composition of each podcast participant so I can easily see the original content along with it's transcript.
While I'm editing the main composition I make sure to highlight sections of the podcast I think would make good clips for sharing across social media websites.
Once I'm ready to start editing on of these clips I highlight the text in the main composition, right click, and clip my selection to a new composition. I put all of these clips in the "Social Clips" folder. Here I can edit the clips down further for the various platforms we advertise the week's episode.
The "Social Clips" folder has another folder within it named "Published". Once I have scheduled one of the clips to be posted (I use later.com), I move the clip into the "Published" sub-folder so I don't accidentally post it again.
When we edit a podcast the last step is to add section markers to the script. These make it super simple to find YouTube clips, any section the is longer than 8 minutes becomes a clip in the YouTube clips folder. I create new compositions of them in the same way that I do for social clips. From the composition I will publish them to the secondary channel directly in Descript.