[mcl22]

By playing around with the german utterances I found something It could have to do with slots not beeing "generic" as you once told me. Or it is a problem with the custom slot examples not always beeing helpful

I already said that Alexa wouldn't recognize the artist "Boney M." or the album "sex sex sex". That's when I use the ListenToArtist with the utterance "ListenToArtist spiel Musik von {Artist}". I use this for Boney M. not for the album of course.
Alexa instead plays Bonnie Tyler who is part of the examples for MUSICARTISTS. The fuzzy matching is not even performed.

If I delete Bonnie Tyler it's getting worse. Trying to make Alexa/Kodi play Bonnie Tyler or Boney M. results in playing Body Count, Bon Jovi ...

But, if use the PlayMedia ("PlayMedia spiel {Movie}") by just saying "spiel Boney M." it works. In this case Alexa gives the string to the skill and the skill tries to match movies, playlists and so on until it matches the right artist.

As far as I understand it, Alexa in this example first looks for a movie title in the samples for MOVIES and can't find any similar enough to match to "Boney M.". So it gives the string to the skill and the skill does the rest. Of course this needs more time because the skill searches for movies, playlists, ...

I ask myself if it's somehow possible to make Alexa do this (just give the string to the skill) in all cases she has no 100% match in the samples? Whatver Alexa does there, the skill would be able to play the right media if Alexa did give the right information