Unfortunatly I have the same experience as you regarding Readarr and audiobooks. I think one of the bigger issues is that audiobooks aren’t indexed a lot on usenet. Both indexera I use get close to no matches when I’m searching for even slightly more niche books (I find epubs all right, but no audiobooks). To solve the issue of finding stuff you would probably need a myanonymouse (private torrent tracker) invite or something, which has focus on audiobooks.
I also find that it syncs my library incorrectly to books (book A as book B, book B also as book B), which then are almost impossible to sort out, and even when it does find books that are monitored I need to manually import them nine times out of ten.
My somewhat best experience was to just DL audiobooks normally and dump them in a ”black hole” directory, and let readarr import them like that.
However, the other issues got to me and I jumped ship.
Hopefully it will become a more mature product in the future, but for now I’m staying to manual handling.
I have a script that I found somewhere and have made personal tweaks to: https://pastebin.com/RJf9kajk Put it in a file called sub-clean.sh and check the instructions in the script on how to add it to bazarr
I have not noticed any issues with it, but there can of course be false positives (like someone saying the word "Facebook" in a series/movie), so your miles may wary. (this is why i leave the .trash.tmp file personally, so that i can restore the removed lines if needed)
and as always, this script is provided as is, but it might help someone out there