Queue length is the instantaneous count of customers waiting to be served. In a phone queue it is the people on hold; in a chat queue it is the open sessions waiting for an agent; in a ticket queue it is the unassigned tickets. A long queue length means a wait time and a likely abandonment.
Supervisors at legacy call centers spend their day watching this number. Spikes drive overtime, holds, and the customer experience deteriorating. The whole staffing math (Erlang C, occupancy targets, shrinkage budgets) exists to keep queue length under a target.
An AI agent collapses the inbound queue. Every call gets picked up immediately, so the queue is the human-handoff queue, which is smaller and slower-moving by design. The metric that matters in an AI-augmented operation is not raw queue length, but handoff backlog at the human team and how fast it drains.
Vorel surfaces the handoff-queue depth in the operator console, with the expected wait and the priority order. The legacy hold queue does not exist because nothing waits to be answered.