A voice application receives no recognized text for a silent audio input. What should its client do?

1. Invent the user's intended spoken command
2. Treat silence as proof of successful authorization
3. Handle the no-result condition and prompt for another valid input where appropriate
4. Execute the last sensitive command automatically