Transcription does not detect all words, when the user answers short phrases such as no, it is not in the transcription, it is Polish, I have tried every model and everywhere there are similar problems. how to fix it? Polish language.

can you please provide the organization ID and the call ID where this occurred?