A few seconds of silence on a call is natural. Five seconds of silence while a customer waits is awkward. Fifteen seconds of dead air while an agent searches for information is a call quality failure. AI detects exactly when and why silence happens.
Dead air refers to extended periods of silence during a call where neither the agent nor the customer is speaking. Unlike natural conversational pauses (1 to 3 seconds), dead air typically lasts 5 seconds or more and signals a problem — the agent is searching for information, navigating a slow system, unsure of the answer, or lost in the conversation flow.
Dead air is distinguished from hold time. When an agent explicitly places a customer on hold with a message, the customer expects silence. When silence happens without explanation during an active conversation, customer satisfaction drops sharply — within as little as 8 to 10 seconds, customers begin to wonder if the call has disconnected.
The customer asks a question the agent cannot answer immediately — product features, pricing details, policy terms. The agent goes silent while searching documentation or a knowledge base. Fix: better training and faster-access knowledge bases.
The agent is waiting for a screen to load, a record to pull up, or a system to respond. The customer hears silence. Fix: technology and process, not coaching — agents cannot speak faster than their systems.
Newer agents frequently fall silent while reading ahead in a script or trying to find the right section of a call guide. Fix: script simplification and script memorisation training.
A customer raises an objection the agent was not expecting and goes silent while thinking of a response. Extended silence here signals the agent needs more objection handling preparation.
Some agents write call notes or update CRM records while still technically on the call, creating silent periods the customer experiences as dead air. Fix: post-call note automation.
In a call transcript, dead air appears as extended gaps between speaker turns — periods where neither the [Agent] nor [Customer] label is active. AI identifies these gaps, measures their duration, and flags instances above the defined threshold (typically 8 seconds) as quality issues.
At scale, dead air data reveals systemic patterns: which agents have the most dead air, at what point in calls it occurs, and whether it correlates with specific call types, objection categories, or product queries. This turns a vague "agents hesitate" impression into specific, actionable data.
Upload any call recording. Get instant timestamped transcript. Identify silence patterns by agent.
Try Free →