A voice interactive session includes detection of an input signaling an interrupt to the session. When the interrupt is detected, instructional and or informational output is interrupted and detection of voice input begins. The voice input is not detected until the output is interrupted. Upon detection of a voice input (or other sound-based input), a determination may be made if the input was valid. If the input was valid, the input is processed, otherwise, instructional and/or informational output may be relayed again and/or the voice input may be redetected.