Does anyone have any recommendations?

The best way, though inefficient, that I can think of is to separate out the session into two. The first to handle separating the individual folders, and the second to handle the separation of the documents. Doing it this way, I can have the first session apply a tag and then have the second session remove the tag from the original source document. By doing this, I create a means by which to have workflow initiated when the processing is complete. I feel this is a poor solution though as it leaves room for error and potentially fail. It also makes the server take longer to process the entire document load, as well as makes it so we spread the resources of the QF server with more sessions running concurrently. Lastly, it is just creating another copy of the original scan, which we then need to do some processing to get rid of it, just so we have the workflows work properly. I would really like to avoid doing a configuration like this, can anyone please advise me with a better solution?