I have a live skill, it's name is "My Audio", and I have a corresponding website https://personalaudio.in/ , in my skill user can play audio form there own files, so user uploads audio in my website and give an invocation names corresponding to each audio file.
I have some utterances configured in my static slot in which I gave some sample utterances like yoga, meditation etc, but user can save any invocation name in my website, so what I have done is I am collecting the invocation names for that user fromm database and pass as dynamic entities.
And finally when user speaks up, I match both static and dynamic entities (slot resolution for both static and dynamic entities).
Is there any better way to implement it? I mean do you think my implementation is fine or is there is any better approach for better matching.
Generally my skill works fine but sometimes users enters arbitary inovacation names like "jen's" or "wohoo" or "b4m" in those cases my entity resolution ( matching of utterances fails)
Please suggest if this approach is fine or I can make it better for better utterance matching.