On the Watch app on your iPhone, go to Messages -> Dictated Messages
Here you'll find options for what to do by default when sending a message to someone. I believe the default is Transcript, which will immediately start listening and convert your message to text. Other options include "Audio" or "Transcript or Audio". It sounds like maybe yours is set to the latter and so it waits for you every time to see which method you want to use? That's my guess anyway.