Queues

Give each queue a purpose

A queue groups work that belongs to a support team, such as access requests, billing questions or product incidents. Set its required skills and assignment strategy around that purpose. The category can suggest a destination, while a general queue provides a visible place for requests that still need classification. Work without a valid destination stays observable instead of silently disappearing from the team’s workload.

A team manager owns staffing and queue health within the assigned scope. Agent identity, team membership and capacity are configured before automatic routing starts. A label saying an agent has a skill is useful only when that person is also active and entitled to work in the destination queue. Administrative access does not permit moving a customer request into an unrelated workspace.

Choose an assignment strategy

Round-robin distributes work through the eligible pool. Least-loaded routing considers the current open workload, while skill matching uses the queue’s stated requirements. Each assignment reserves capacity as it selects an owner. Two requests arriving together must not both rely on the same last available slot and exceed the agent’s configured limit.

When no eligible agent can accept the request, leave it unassigned with an understandable reason. The manager can adjust staffing or choose a permitted overflow destination. An overflow chain must terminate safely; routing a ticket repeatedly between two full queues does not create capacity or establish responsibility. A failed automatic assignment remains something the team can act on.

Make reassignment a real handoff

Suppose an L1 agent discovers that a ticket needs a product specialist. A permitted reassignment identifies the new agent and records why responsibility changed. The receiving person sees the current customer context, due targets and unresolved question. The earlier owner remains in assignment history so a later workload review can distinguish original intake from specialist work.

The same principle applies when someone becomes unavailable during the day. A manager can move authorized work within the team without rewriting who handled it earlier. Assignment history records the actor or the named routing service rather than inventing a human agent for an automated action. Retrying the accepted reassignment returns its original result.

Review the working record

Review the unassigned queue at each staffing handover. An empty assignee is a visible operational exception, not an invitation to let a customer choose an internal owner. Resolve the reason for missing capacity before retrying the same routing command.