Call Center Outsourced evidence brief · Desk review · Published
How should a contact center keep a queue running during a floor evacuation?
An evacuation procedure that names the trigger, preserves live calls, assigns floor roles, and reconciles the queue before reopening.

On this page
Key stats
- 2.67 million U.S. customer service representative jobs in 2025 (BLS)
- $21.53 median hourly wage in May 2025 (BLS)
- 289,500 projected annual openings, 2025 to 2035 (BLS)
Key takeaways
- Give the site safety lead authority to evacuate without client approval.
- Test the carrier reroute before an emergency, not during one.
- Name a floor warden, a headcount owner, and a communication owner.
- Reconcile dropped calls and open promises before the next shift.
Question and scope
This brief asks how a contact center should keep a queue running during a floor evacuation. The population is a site with agents on live calls and a building alarm that requires them to leave. The unit is one evacuation event, from the trigger to the queue's recovery. The analysis covers authority, call preservation, floor roles, and the record. A short card at each station is more useful than a long plan in a drawer.
The scope excludes the building's fire-safety design, which belongs to the landlord and local authorities. It also excludes a full disaster recovery plan for a destroyed site. The queue still needs a short evacuation procedure because a live call cannot be paused like a document. Write the procedure with the client and the site safety lead. The procedure should be tested before an alarm tests it.
What the evidence supports
CISA continuity guidance treats an incident as an event with roles, communication, and recovery steps. ISO 18295 places contact-center operations within a managed system. NIST log guidance supports reconstructing what happened after the event. None of these sources defines an evacuation trigger or a safe call-handling rule for every building. Local fire codes and the site's safety plan set the legal minimum.
The evidence supports preparation over improvisation. A team that knows who calls the alarm, who speaks to customers, and who accounts for staff can leave faster. A team that debates authority during the alarm loses time. Record the decisions in a one-page card at each supervisor station, not only in a binder. The card should name backups for every role.
A useful test is a short drill that starts with a live call and ends at the assembly point. Time the call message, the exit, and the headcount. The result shows which step needs practice, not whether the team is brave. Repeat the drill when a floor layout or a supervisor changes. The drill should not use a real customer.
Define the trigger and the authority
Name the trigger that requires an evacuation. The building alarm is the obvious one, but a chemical spill, a threat, or a gas smell may require the same response. The site safety lead or the highest-ranking manager on duty should have authority to order the evacuation. Do not require a client approval before people leave. Safety decisions belong to the site, not the account team.
Tell agents what to do with a live call when the alarm sounds. The approved script may ask the customer to hold, transfer to a recorded message, or end the call with a callback. The choice should depend on the system's capability and the time available. Agents should not improvise a promise during an evacuation. A short phrase is easier to say under stress than a long explanation.
Plan the call-preservation step
Decide in advance whether the telephony system can route calls to a backup site or a recorded message. Test the transfer before an emergency, not during one. If the system cannot reroute, the queue may need to stop new calls at the carrier level. Name the person who can make that change and the number to reach them. The carrier contact should be available outside business hours.
Preserve the active call records before the site goes dark. Confirm that the platform writes notes and recordings to a remote store. Do not rely on an agent's desktop saving a draft; If the system fails, record the event on paper and reconcile it later. The customer's request should survive the evacuation even when the building does not. A paper log with times and case numbers is enough for reconciliation.
Check whether the recorded message tells the customer what will happen next. A message that only says the office is closed leaves the caller with no path. Offer a callback number or a website that works during the outage. Test the message from an outside line, not only from a desk phone. The customer should hear the same option the agent would offer.
Assign floor roles
Assign a floor warden, a headcount owner, and a customer-communication owner. The warden leads the exit and checks the route; the headcount owner accounts for every person at the assembly point. The communication owner updates the client and the carrier. One person should not carry all three roles in a large queue. Name a backup for each role and put the names on the card.
Give supervisors a small grab list: the evacuation card, a charged phone, and the carrier contact. Do not ask them to collect laptops or files. The headcount is more important than the equipment. At the assembly point, report the count and any missing person to the warden. Keep the report factual and brief; A missing person is an emergency, not a scheduling problem.
Recover the queue after re-entry
Re-entry should be authorized by the site safety lead, not by the queue manager alone. Before agents return, check power, network, and the telephony link. Test one call path before opening the full queue; A site that looks normal can still have a failed system. Record the test result and the time the queue reopened. Do not reopen to clear a backlog if the building is not safe.
Bring the queue back in stages; start with one team and a small call volume. Check the recording, the notes, and the routing before adding more agents. If the carrier rerouted calls, confirm that the route is back. Tell the client what is live and what is still limited. A staged recovery is easier to reverse than a full restart; the first hour should be treated as a control test.
Decide which team returns first and why. The team that handles the oldest promises may not be the team that answers new calls. Write the order in the recovery card with a named supervisor for each step. A staged return without an order becomes a race to the first free desk. The order should follow customer impact, not the loudest queue.
Reconcile the record
Reconcile the calls that were active when the evacuation began. Match the platform record, the paper log, and the customer's account. Identify calls that dropped without a callback and assign an owner. Do not assume that a dropped call was abandoned by the customer. The event may have cut the line; A short list of open promises should go to the next shift.
Record the evacuation time, the re-entry time, and the queue status for each interval. Note the calls rerouted, the calls lost, and the callbacks created. Keep the record with the incident, not in a general shift note. The client will ask what happened to its customers, and the timeline is the answer. The timeline should distinguish a system failure from a safety delay.
Limits and review
This procedure cannot guarantee that every customer call survives an evacuation. It also cannot replace the building's fire-safety plan or local law. The queue can control the trigger, the roles, the reroute test, and the reconciliation. State those limits in the client review so no one treats the plan as a guarantee. Safety always outranks the service level; the client should agree to that priority in writing.
Review the procedure after every evacuation and at least once a year. Test the carrier reroute and the headcount process; update the card when a supervisor or a system changes. Sample one drill for the time from alarm to customer message. A plan that has never been drilled is only a document. The drill should include a live call, not only a meeting.
Keep the drill record with the client's continuity file. Note the date, the scenario, the time to evacuate, and the time to reopen. List the fixes and the owner for each. A record that shows only a successful drill without a fix is not useful for the next review. The next drill should test the fix in the same conditions that exposed it.
Put this into a support lane
Bring the current evacuation card, the carrier contact list, and one drill record so the review can test the trigger and the recovery steps.
Review the evacuation cardRelated operating guides
FAQs
Should agents stay on live calls during an alarm?
No. Safety comes first. The approved script should move the call to a recorded message, a backup site, or a callback before the agent leaves.
Who authorizes re-entry?
The site safety lead or the authority named in the building's safety plan, not the queue manager alone. The queue should test systems before reopening.