Hello! Is there anything in your roadmap about enhancing the accuracy of the voice to text transcript feature? Do you perform recurring training for the speech recognition or is it speaker independent? More often than not, our transcripts lacks accuracy and is not very dependable without the audio to confirm what was said. Even when words are clearly spoken and audible to me, the listener, they are not correctly transcribed. This makes it harder to rely on the really neat "global transcript search" feature because if the word wasn't transcribed accurately, the search will return no results.
Moreover, we are a unique industry and wonder if there are certain vocabulary words that could be feed into this feature so it could correctly hear what was said. An example, the caller said "terramation", but it was transcribed as "termination".
Curious if others struggle with the speech recognition being accurate?
