What is voice activity detection (VAD)?
Technology that detects when someone is speaking and when they've stopped.
Voice activity detection (VAD), explained
VAD tells a voice agent when the caller has started talking, so it can stop speaking, and when they have finished, so it can reply.
Tuning VAD is a trade-off: too eager and the agent interrupts pauses; too slow and the conversation feels laggy.
On noisy phone lines, from traffic, televisions or crowded shops, VAD has to tell speech apart from background sound so the agent doesn't stop talking for every noise.
See how these ideas come together in Chitti, Manolle's AI voice agent.
Related terms
- Barge-inThe caller's ability to interrupt the voice agent while it is speaking, and have it stop and listen.
- Turn detection (endpointing)Deciding when the caller has finished their turn so the agent can reply.
- Latency (voice response time)The delay between the caller finishing speaking and hearing the agent's reply.
See an AI voice agent at work on your calls.
See the complete system live: marketing, lead capture, AI voice calling, sales conversion, and analytics.