It's also much quicker to parse through text than to wait for someone to say it, even if they are a fast speaker.