While not the first app of its kind, it does raise the question: if these start-ups can do speech to text succesfully, why is Google's voicemail transcription a

While not the first app of its kind, it does raise the question: if these start-ups can do speech to text succesfully, why is Google's voicemail transcription and automatic subtitling so inaccurate? Serious question. Is it the quality of the input and an inability to adapt to changing bandwidth?