Record from the car microphone has the details on how to implement voice input in cars. In the sample flows on the page you linked, you as the developer are responsible for beginning to record audio when the user taps the button, as well as for processing the audio and updating the UI (e.g. showing the toast in the final screenshot).

The only UI/functionality that is handled for you is showing the indicator that audio input is being capture.