What it does

Defines what happens to a work item when a queue can't serve it, queue full, outside operating hours, or predicted wait too long. Two trigger points: pre-queue (before the item enters the queue) and post-queue (while it's already waiting).

Key facts

• Pre-queue conditions: Out of operating hours, Work item limit exceeds (1–100), Average wait time threshold (voice only, 30 sec–60 min)
• Post-queue conditions: Wait time in queue exceeds (voice: 1–60 min; messaging: 1 min–2 days; records: 3 min–2 days)
• Voice pre-queue actions: Assign anyway, Direct callback, End call, Transfer to queue, Transfer to external number, Voicemail
• Messaging pre-queue actions: Assign anyway, End conversation, Transfer to queue
• Post-queue voice actions (transferred/overflowed items): Transfer to queue and Transfer to external number only: callback/voicemail not available for these
• Pre-queue check runs once only: if supervisor transfers after overflow fires, no further pre-queue checks occur
• Fallback queue always bypasses overflow: items routed to fallback are always accepted
• Queue count for work item limit condition updates every 10 seconds: brief burst overflow possible near threshold
• Rule-specific overflow overrides can be set per route-to-queue rule: overrides queue-level defaults for specific work items (e.g. VIP work never ends a call, even if the queue's default is "End call")
• AI-powered immediate overflow (Preview): fires instantly when no eligible rep is available, without time-based conditions; configured via Conversation Orchestration playbooks

When to use / skip

Always configure overflow for production deployments. Without it, customers wait indefinitely when queues are full or closed. Direct callback is the highest-value voice overflow action for most contact centres.

Configuration decisions

• Pre-queue vs post-queue conditions: both serve different scenarios; pre-queue handles arrival-time decisions, post-queue handles already-waiting items that breach SLA
• Direct callback vs voicemail vs transfer: callback is best for voice when reps will be available soon; voicemail for after-hours; transfer to another queue for cross-team coverage
• Rule-specific overrides: use for tiered service (VIP customers bypass the same overflow actions that apply to standard customers on the same queue)

Gotchas

• Operating hours on a queue auto-creates an overflow condition that can't be deleted: only its action can be changed.
• Out-of-hours overflow must be explicitly tested in UAT. Testers work during business hours and almost never test this. Deliberately set a queue to "out of hours" during testing.
• Voice post-queue actions are restricted for transferred/overflowed items. A call that's already been transferred or overflowed once can only be transferred again or sent to an external number: not callback or voicemail. Plan multi-hop scenarios accordingly.
• AI-powered immediate overflow is Preview. Don't use in production for regulated environments yet.

Consultant notes

• Out-of-hours overflow is the UAT test that almost never gets run because testing happens during business hours. Build it explicitly into the UAT test plan and force a test by temporarily adjusting operating hours to simulate closed. "We'll test it after go-live" isn't acceptable for an overflow path that goes straight to End call or Voicemail.
• Rule-specific overrides for VIP or priority customers are worth raising in requirements if the client has any tiered service model. It's a clean solution that avoids the overhead of a separate VIP queue just to get different overflow behaviour.
• Direct callback is usually the highest-value voice overflow action for most contact centres. If the client defaults to "End call" because it seems simpler, walk through the abandonment and CSAT impact: the conversation usually changes.

---

Source last updated: 2026-04-27 | Review when: AI-powered immediate overflow reaches GA or new overflow action types added