Voice Activity Detection (VAD) is a technology that detects whether an audio stream contains human speech or silence. It separates spoken audio from background noise and non-speech sounds.
Voice AI platforms use VAD to determine when users begin and finish speaking, reduce unnecessary speech processing, improve turn-taking, lower latency, and optimize speech recognition performance.