1. Establish the operating rule: one roster, separate decisions
Move responsibility for each active visit—not just its data. For a warehouse visitor management system migration, use one authoritative operational roster throughout the changeover, and route every arrival, departure and correction through a named owner.
This procedure assumes a planned replacement, functioning systems and drivers or contractors whose visits span the changeover. Historical-record migration, procurement, general implementation and network-outage recovery are outside scope. The operating controls below are recommendations for local approval, not universal legal requirements or promises about software capabilities.
Keep four decisions separate throughout:
- Visit record: which system holds the operational record and its original arrival time.
- Approval: who authorised the visit, its scope and whether that authority remains valid.
- Credential: which badge or electronic permission is currently valid, restricted, expired or revoked.
- Presence: what is actually known about the person’s arrival, location and departure.
Do not treat a transferred record as automatic transfer of approval, credential validity or evidence of physical presence. Require separate verification of each. Do not ask people already on site to sign in again merely to populate the replacement system.
AWS’s migration guidance recommends named task owners, sequenced activities, rehearsals and a documented rollback procedure. Those are useful cutover disciplines here; its guidance concerns technology migrations, not warehouse admission or safety law. AWS: pre-cutover stage, checked 19 November 2026
2. Assign owners and define the authority boundary
Recommended preparation: appoint named people and deputies covering the entire change window and the following shift.
| Role | Responsibility during changeover |
|---|---|
| Cutover lead | Authorise the start, pause, final handover or rollback; resolve conflicting instructions. |
| Outgoing system administrator | Produce the baseline, explain old-system behaviour, control its remaining writes and preserve source records. |
| Incoming system administrator | Demonstrate supported transfer methods, maintain identifier mappings and verify target records and departures. |
| Gatehouse supervisors | Control transactions at every entrance and exit, brief operators and acknowledge changes of authority. |
| Roster custodian and deputy | Maintain the single consolidated operational roster, reconcile events and hand over ownership between shifts. Name a supervisor or administrator explicitly to this role. |
| Access-control owner | Approve and execute credential and integration changes under the site’s access-control procedure. |
Agree a cutover timestamp, T, including the date, time zone and UTC offset. At T, responsibility for recording new arrivals changes. Record the later acceptance time separately: starting the changeover is not the same as completing it.
Use these authority boundaries:
| Stage | Authoritative operational roster | Permitted transaction route |
|---|---|---|
| Before T | Old-system roster, reconciled for known exceptions | Existing approved workflow. |
| From T through verification | Access-restricted consolidated roster owned by the custodian | Each event goes through the controlled transition process; administrators update only the assigned system records. |
| After successful complete transfer | New-system roster, once formally accepted | New system only; old records are non-operational. |
| While any old-cohort visit remains outside the new system | Consolidated roster | Old-cohort departures use their assigned old records; new arrivals use the new system. |
Give authorised gatehouse and emergency-accountability staff access to the current roster, its update time and its owner. Restrict editing to the custodian and explicitly delegated staff. Avoid distributing independently edited copies.
At each boundary, obtain acknowledgement from every gatehouse supervisor. Do not permit staff, hosts, kiosks or other enabled routes to independently admit or sign out the same people in both systems.
3. Verify what can actually transfer
Do not infer active-visit migration from a CSV import feature. SwipedOn’s documentation dated 19 May 2026 describes importing visitors for bulk preregistration, with optional invitations and rejection handling. It does not establish an active-visit transfer method. SwipedOn: bulk preregistration
Similarly, Envoy’s export documentation dated 12 October 2026 lists sign-in and sign-out information and other selectable fields. An export capability alone does not demonstrate that another system can reconstruct those visits or their history. Envoy: exporting visitor data
Recommended verification: require documentation and a witnessed test for the installed versions, subscription and integrations. Record the method, configuration, result and vendor-confirmation reference—not simply “migration supported.”
| Check | What to demonstrate before using live visits |
|---|---|
| Supported export and import | The approved interface, included fields, rejection reporting and retry behaviour. Establish whether it creates profiles, invitations, completed visits or genuinely active visits. |
| Original arrival | Preserve the actual arrival date, time and time zone separately from import or creation time. Check cross-midnight visits and unknown source times. |
| Active status | The person appears in the operational roster and can subsequently depart through the intended route. |
| Unique identifiers | Retrieve stable old and new visit identifiers; distinguish a person, a recurring profile and a particular visit. |
| History and approval | Establish what remains in the source, what transfers and what cannot be represented. Do not assume audit history or approval evidence transfers. |
| Notifications | Demonstrate whether import, creation, correction or retry sends host alerts, invitations or other messages. Suppress migration-generated duplicates using supported controls. |
| Credentials and integrations | Establish which actions create, change, expire or revoke access; test the intended ownership handover. |
| Automatic sign-out | Inspect scheduled clearing, maximum-duration rules and any overnight behaviour in both systems. Demonstrate how spanning visits remain accountable. |
| Departure and corrections | Complete a departure and an arrival-time correction without a second admission, duplicate count or unintended access grant. |
Treat unsupported or undocumented behaviour as unknown—vendor confirmation required. Written assurance should identify the exact supported method; test its operational result before relying on it.
For example, Envoy’s sign-out documentation dated 23 May 2025 describes scheduled automatic sign-out. Its entry-editing documentation dated 11 August 2026 allows arrival-time edits but identifies approval history and integration status among automatically recorded, non-editable information. These distinctions reinforce the need to test status, timestamps and history separately. Envoy: sign-out behaviour · Envoy: editing entries
4. Select the transfer route and preserve credential ownership
Recommended decision: use direct active-visit transfer only when the complete path has been demonstrated. Otherwise, finish existing visits in the old system while sending new arrivals to the new one.
Direct transfer
Prepare a cross-reference connecting the source system, site and old visit identifier to the new visit identifier. Maintain one operational entry per person’s current visit, even where both systems retain a record. Confirm the person using the site-approved identity method and relevant visit details. Never use a shared vehicle registration as the person’s identity or merge passengers with a driver.
Carry the original arrival time and its provenance separately from migration time. If the target cannot represent the original arrival accurately without falsifying a new admission, use the old-cohort route instead. Retain unknown times as unknown.
Preserve references to approval evidence and the responsible host. Do not manufacture a target approval history or mark documents signed because a source row says the person was admitted. The responsible approval owner should resolve any gap before affected access or activity continues.
Keep migrated source records non-operational through the supported, tested method. Do not bulk-sign people out merely to clear the old dashboard. If a supported retirement method would falsely imply departure or disrupt credentials, retain consolidated authority until that issue is resolved.
Controlled old-cohort completion
At T, stop new admissions into the old system. Keep each pre-cutover visit assigned to its old record until confirmed departure. Process new arrivals only in the new system.
The named custodian maintains one access-restricted consolidated roster containing both cohorts, original arrivals, current status, assigned departure system and unresolved exceptions. An old-system list beside a new-system list is not an accepted substitute: reconcile them into one deduplicated operational view.
Keep old-system staff access and necessary credential dependencies available until the last old-cohort visit has departed. At shift handover, identify every remaining old visit and who will close it. Do not time-limit the cohort by inventing departures.
Credential continuity
Have the access-control owner verify existing credential assignments, permitted zones, expiry and revocation status independently of visit transfer. Preserve existing restrictions and expiry rather than granting a fresh allowance at migration time.
Integration effects are configuration-specific. For instance, Envoy’s Brivo documentation dated 12 August 2026 describes visitor access triggered by invitations or sign-ins and licence-plate credential removal at sign-out or expiry. Do not apply that behaviour to another integration without confirmation. Envoy: Brivo integration
5. Rehearse the transition and make the go/no-go decision
Recommended rehearsal: use fictional visitors in a controlled test environment with the intended configuration. Exercise an overnight visit, a rejected import, a duplicate, a departure during transfer, a correction, an incompatible badge and rollback after a new arrival.
Test both the normal route and old-cohort completion. Have gatehouse staff locate a person, establish the departure owner and complete sign-out from the same roster they will use during changeover. Test notification suppression for migrated visits without suppressing required notifications for genuinely new arrivals.
AWS recommends functional and integration testing, operational readiness review and predefined rollback checkpoints. AWS: cutover stage, checked 19 November 2026
Recommended go/no-go gates: the cutover lead should require demonstrated evidence for every gate below, first at rehearsal and again against live conditions immediately before T.
- Roster completeness: every current visit and unresolved presence is represented once across all covered entrances; no unexplained omission or duplicate remains.
- Departure handling: staff can close a visit from either cohort, record actual departure and verify the assigned credential action.
- Credential continuity: the access-control owner accepts the tested route, including restrictions, expiry and preservation of revocations.
- Roster access: every duty gatehouse and designated emergency-accountability role can retrieve the authoritative view and identify its freshness.
- Transaction control: operators know the timestamp boundary, event route and assigned system; the team can reconcile retries and corrections.
A failed direct-import test need not block an independently tested old-cohort route. A failure to maintain a trustworthy consolidated roster does block the changeover. Do not start simply because the scheduled time has arrived.
6. Execute the cutover and apply transition events exactly once
Recommended live sequence: preserve the baseline, route changes through one controlled event record and reconcile each event before accepting the target roster.
- Reconcile immediately before T. The outgoing administrator and gatehouse supervisors confirm all entrances, active visits, cross-midnight records and known discrepancies. Preserve a timestamped baseline. Keep uncertain departures explicitly unresolved, with a last reliable location or observation and a named investigator.
- Activate consolidated authority at T. The lead announces the change and obtains gatehouse acknowledgements. The custodian begins maintaining the baseline plus transition events as the authoritative operational roster. Restrict ordinary old-system writes as rehearsed, while allowing assigned old-cohort departures where applicable.
- Capture each event once. Give each arrival, departure or correction a unique transition event identifier. Record when it happened, when it was reported, the person and visit reference, reporting operator, assigned system and processing outcome. Where the event time is unknown, preserve that uncertainty.
- Apply the event through its assigned owner. New arrivals follow the new system’s normal approval and admission route. Departures update the currently assigned visit record. Corrections retain the prior value, corrected value, reason, evidence and authoriser without representing another physical admission.
- Acknowledge application. Record the resulting system identifier or action reference and who verified it. Before retrying an ambiguous submission, inspect the destination for an existing result. Do not create another event identifier or resubmit blindly.
- Reconcile at a stated checkpoint. Match every baseline visit to its target record, old-cohort continuation or confirmed departure. Match every transition event to a verified result or an explicitly owned pending action. Continue capturing later events while checking the bounded set.
Treat “exactly once” as an acceptance outcome to verify—not an assumed property of imports or APIs. Separate the logical event from its technical processing: one confirmed departure may require coordinated updates to linked records, but it must reduce the operational count only once.
For a departure during transfer, update the authoritative roster immediately from the confirmed event. If no target record exists, prevent a later import from resurrecting the person as active. If a target record already exists, close that linked record using the same event and actual departure time. Do not ask the person to sign out again.
Before final acceptance, reconcile identities as well as totals. Keep unresolved people visible and report their number separately from confirmed presence; a matching total does not resolve an identity mismatch.

7. Pause or roll back without losing people or transactions
Recommended exception branches: the cutover lead decides whether to continue under a tested alternative, pause for correction or roll back.
| Exception | Immediate response | Condition for continuing |
|---|---|---|
| Active-visit import rejected | Keep the person on the consolidated roster under the old visit identifier. Preserve the rejection reason. Check whether any partial target record exists before retrying. | A supported correction succeeds, or the lead assigns the visit to tested old-cohort completion with a named departure owner. |
| Duplicate records | Stop retries. Verify identity and visit linkage; count the logical visit once while retaining both references for investigation. | Administrator resolves the duplicate through a supported method without removing a different person or triggering unintended credential actions. |
| Departure during transfer | Capture one confirmed departure event; mark its operational effect immediately and reconcile any target record created by the transfer. | Both record handling and access removal are verified, with no later import able to reactivate the visit. |
| Incompatible badge | Keep the person’s presence recorded. Refer replacement or interim escort arrangements to the access-control owner. | Approved credential transition is completed and checked. Do not extend access or restore an invalid credential merely to make the changeover work. |
| Untrustworthy consolidated roster | Pause the changeover and new admissions; continue departure reporting and presence investigation. Notify the duty operational owner. | A complete, accessible operational roster is re-established and accepted. Do not treat either partial dashboard as authoritative. |
Roll back without discarding new transactions
AWS distinguishes rollback before data changes from rollback after new transactions, when the former source may be stale. AWS: rollback with changed data
Recommended rollback sequence:
- The lead declares a rollback boundary and keeps the consolidated roster authoritative. Pause new admissions if transaction control cannot be maintained; do not obstruct safe departure.
- Preserve both current system states, the baseline, identifier mappings and all transition events through that boundary. Continue logging subsequent events through one controlled route.
- Reconcile every new arrival, departure, correction and credential action since T. Do not restore a pre-T snapshot and discard those changes.
- Use the rehearsed supported method to represent new arrivals in the old system, with original arrival times and no duplicate admission effects. If that is not supported, leave those visits in the new system as a separately assigned cohort and retain consolidated authority.
- Apply confirmed departures and corrections to the restored operational route once. Check that replay or integration restart cannot recreate access for departed people.
- Have the access-control owner reconcile credentials against current approved decisions, including every revocation since T. Never restore revoked credentials from an old configuration or snapshot.
- Verify the same completeness, departure, credential and staff-access gates before announcing the revised authority boundary.
If connectivity fails, activate the separate warehouse network-outage procedure rather than improvising recovery inside the migration. If an emergency occurs, stop migration work as directed by the site plan and give the accountability lead the current roster, its timestamp and unresolved entries; use the evacuation roll-call reconciliation procedure.
8. Worked cutover: 12 visits become 13, with one rejected import
Fully hypothetical completed example: all names, identifiers, times, outcomes and technical methods below are invented. They describe no real product. The fictional warehouse has rehearsed direct transfer and an old-cohort exception route, including independent credential handling.
The cutover is 20 November 2026 at 02:00 GMT, with acceptance at 02:15 GMT. Lead Maya authorises decisions; outgoing administrator Owen controls old records; incoming administrator Imani controls target records; gatehouse supervisor and roster custodian Lee maintains consolidated authority; access owner Arun controls credentials.
At 01:59, Lee and Owen reconcile 12 active visits, O-101 through O-112, with no unresolved presence. They retain the baseline and original arrivals. Contractor Jo, O-112, arrived at 20:40 GMT on 19 November and remains with the maintenance host.
| Time | Event and completed handling | Authoritative active count |
|---|---|---|
| 02:00 | T takes effect. Lee’s restricted consolidated roster becomes authoritative; new arrivals are routed to the new system. | 12 |
| 02:02 | Eleven visits, O-101–O-111, transfer to N-201–N-211. Jo’s O-112 import is rejected because its category has no accepted target mapping. Lee retains Jo as active, assigned to old-system departure handling. No new arrival is recorded. | 12 |
| 02:03 | Event E-001: driver O-101/N-201 departs through the gatehouse. Lee records the confirmed departure; Imani closes the target record at 02:03. Arun verifies access removal. | 11 |
| 02:05 | Event E-002: new driver Priya arrives, passes normal checks and is admitted as N-301. The normal host notification is sent once. | 12 |
| 02:07 | Event E-003: contractor O-102/N-202 departs. The same controlled departure process completes, including credential removal. | 11 |
| 02:08 | Event E-004: new contractor Luis is admitted as N-302 after the required checks and acceptance. | 12 |
| 02:09 | Imani confirms no partial target record exists for Jo. Maya approves old-cohort completion instead of another import attempt. Jo’s host confirms continued presence; Arun confirms existing restricted access remains valid until its existing expiry. | 12 |
| 02:10 | Event E-005: new visitor Erin is admitted as N-303 through the normal workflow. | 13 |
| 02:11 | Event E-006: a verified host-name correction is applied to N-206 with its previous value retained in the correction record. It creates neither an admission nor an occupancy change. | 13 |
| 02:15 | Lee and both administrators reconcile the baseline, mappings and all six events. Maya accepts operation under the consolidated-roster route, not new-system-only authority. | 13 |
Final reconciliation: 12 starting visits − 2 confirmed departures + 3 arrivals = 13 active visits.
The 13 comprise nine transferred visits still active in the new system, three new arrivals active in the new system and Jo’s one old-system visit. The new system therefore shows 12 active visits; the consolidated roster shows 13. The old records for transferred people remain preserved but non-operational under the rehearsed method, with no invented departure timestamps.
Jo’s import rejection is operationally resolved by a named continuation route—not by making Jo disappear from the target count. Jo retains the original 20:40 arrival; the rejected attempt is recorded separately at 02:02. Owen owns the eventual old-system sign-out, and Lee owns roster inclusion until that departure is confirmed.
At acceptance, the team retains the baseline, eleven identifier mappings, rejection result, six applied event acknowledgements, two verified credential removals and the roster showing no unresolved presence. Transferred visits generated no repeat host notifications or second admissions. Imani owns the category-mapping defect, with review at 09:00; Owen and Lee hand Jo’s remaining visit explicitly to their next-shift deputies. The old system is not yet retired.
- Starting visits 12 At 02:00 GMT, the consolidated roster takes authority over all pre-cutover active visits.
- Transfer outcome 11 + 1 Eleven visits transfer; Jo’s rejected import remains active under the old-cohort departure route.
- Departures −2 Confirmed departures at 02:03 and 02:07 are applied once, with credential removal verified.
- Arrivals +3 New arrivals at 02:05, 02:08 and 02:10 use the new system’s normal admission workflow.
- Reconciled active visits 13 At 02:15, twelve active visits are in the new system and one remains in the old system. The consolidated roster remains authoritative.
9. Close with evidence, named exceptions and a separate retirement decision
Recommended closure: separate acceptance of the new arrival workflow, completion of active-visit transfer and retirement of the old service. Record the timestamp and decision owner for each milestone.
The cutover lead should accept the following evidence before ending consolidated authority:
- A reconciled baseline and final roster, with each person traced to an active visit or evidenced departure.
- Old-to-new identifier mappings, original arrivals and separate migration timestamps.
- Every transition event accounted for, including retries, corrections and rejected or partially processed records.
- Verified departure handling and credential continuity, including preservation of revocations.
- Confirmation that migration did not produce duplicate admissions or unwanted repeat notifications.
- No remaining old-cohort active visits; any uncertain presence remains visible with an investigator, next action and escalation time.
- Acknowledged staff access to the accepted roster and a completed shift handover.
If a defect remains, name its owner, operational restriction, next review and closure evidence. Keep consolidated authority whenever a remaining exception prevents the new system from being the complete operational roster. Do not relabel a material gap as a minor follow-up merely to close the project.
Once the old cohort is finished and dependencies are resolved, the outgoing administrator can retire operational access under the approved plan. The access-control owner should separately approve removal of obsolete integrations and confirm that required current access remains intact.
Retiring the old system is not permission to delete its records. Have the records owner govern the retained source records, mappings, transition evidence and temporary exports under the existing policy and preservation instructions. Use the separate warehouse visitor-log retention procedure for authorised disposal rather than deleting records as a migration-cleanup shortcut.
Conclusion
Make the handover complete at the person level: preserve the original arrival, verify continuing approval and credentials, assign one departure route and keep every unresolved visit visible. If active visits cannot transfer reliably, retain the old cohort under one consolidated roster rather than forcing a second sign-in. Close only when transactions reconcile and remaining responsibilities are explicitly accepted.
Sources
Source pages checked 19 November 2026.
Pricing and plan details can change; verify current terms with the vendor.
