I use zotfile (http://zotfile.com) to handle this.
Files are all stored in a single folder (which can be in different locations depending on computer) which I point zotfile to. Inside that each paper is put into a folder based on author names. And each pdf is then renamed based on authors, year, and title.
All of the above is customizable and automated.
But in some ways Zotero 6 makes filesystem workflows more complicated, in that PDF annotations don't get stored as actual annotations in the PDFs.
It's been working great for me, although recent versions have a problem that it no longer recognizes the root directory and pollutes the zotero database with machine-specific absolute paths ((
This is great to have consistent filenames from lots of different sources. Of course if the metadata is wrong or missing, it won't help.