just made a breakthrough - changing the string from KUSG%C3%96Z to KUSG\xC3\x96Z, and using the latin-1 encoding seems to do the conversion I need.
Leave the posting, hopefully it might spare some one else a couple hours of head banging.

I've written a python function that retrieves data from a sqlite (Banshee - Linux music player's) database, ultimately ending up with a path/file name. That name appears to be in an "enhanced ASCI" hybrid encoding, using %20 for spaces, and, as an example KUSG%C3%96Z - %C3%96 is a capital O umlaut (the string is KUSGÖZ). I'm using the path/file to copy files from one location to another. Replacing %20 with a space works for path/files...