That's Strange, I'm testing a handfull of conversational voice ai's and they all have the ability to do this, why is this not the case here?