Moving from one MDM or UEM platform to another means every laptop, phone and tablet briefly has two managers or none, and that gap is where devices fall out of compliance, lose encryption escrow, or keep trusting a console you meant to switch off. A device management migration is not a data copy: management profiles, certificates and enrolment records cannot simply be exported and imported, so in practice every device has to be unenrolled from the old platform and enrolled in the new one, while access policies keep working. This vendor-neutral playbook is for endpoint and workplace engineering teams, identity teams that rely on device compliance for conditional access, and security leads who must sign off the move. It covers the inventory, the parallel-run design, the device posture gap, enrolment waves, platform credentials, rollback and closing the old platform. How to secure a device fleet in steady state is in the Mobile Device Security Guide, and a first deployment is covered in the MDM and UEM platform rollout playbook.
Key takeaways
- Plan for re-enrolment, not migration. Treat every device as a fresh enrolment into a new trust relationship, and design the waves, user communications and support capacity around that.
- Keep conditional access honest during the overlap. Identity policies must accept a compliance signal from either platform for a defined window, and never fall back to accepting unmanaged devices because the signal is missing.
- Move the recovery material first. Disk encryption recovery keys, activation lock bypass codes and similar escrowed secrets must be confirmed in the new platform or a records store before a device leaves the old one.
- The old console stays dangerous until it is gone. It can still wipe, lock or push software to any device it manages, so reduce its administrators to a named few on day one and decommission it deliberately.
- Measure the gap, not just the finish line. Track devices managed by neither platform every day of the migration; that number, not the percentage moved, is the risk.
Why device management migrations are different
NIST SP 800-124 Rev. 2, Guidelines for Managing the Security of Mobile Devices in the Enterprise, frames device management around the lifecycle phases of deployment, use and disposal, for both organisation-provided and personally owned devices. A platform migration forces every device back through the deployment phase while it is in use. The NCSC notes that a device management service holds the keys to unlock, decrypt and wipe devices, so it needs strong protection itself. During a migration you run two such services at once.
Three properties make this harder than migrating most enterprise platforms:
- State lives on the device: configuration profiles, Wi-Fi and VPN certificates, managed applications and their data containers are installed by the old platform and are usually removed when it unenrols the device. The user can lose network access and work data at the moment of transfer.
- Automated enrolment is bound to one server: corporate devices registered in the device maker's automated enrolment programme are assigned to a specific management server. That assignment must be changed in the programme's portal, and devices typically only pick up the new assignment after a factory reset or at next setup.
- Access decisions depend on it: where the identity provider checks device compliance before granting access, a device mid-migration has no current compliance state. Get the policy wrong and users are locked out, or worse, everyone is let in.
Inventory: what has to move
Start with an inventory owned jointly by the endpoint and security teams. Build the device denominator from purchase records and the asset register, not from the old platform's own count, so that devices that already dropped out of management are found now rather than after cutover.
| Object | Security concern | Typical decision |
|---|---|---|
| Device records and ownership type | Unknown or stale devices carried across; personally owned devices treated as corporate | Reconcile against the asset register; retire stale records instead of moving them |
| Disk encryption recovery keys and lock bypass codes | Devices become unrecoverable, or keys are left in a console that is later deleted | Escrow in the new platform at enrolment; export and archive the old set under restricted access until devices are confirmed |
| Compliance and configuration policies | Settings silently weaker in the new platform (passcode length, encryption, OS version floor) | Rebuild from a documented baseline and compare setting by setting |
| Certificates and network profiles | Wi-Fi, VPN and authentication certificates issued through the old platform's connector | Issue new certificates through the new connector; revoke old ones as devices move |
| Managed applications and app protection policies | Work data in managed containers lost or left unprotected on personally owned devices | Rebuild app protection; plan data sync before unenrolment |
| Administrator roles and scopes | Broad build-time roles never reduced; old administrators keep wipe rights | Recreate from current job functions; freeze the old platform's role list |
| Integrations and API credentials | Identity provider, certificate authority, directory, ticketing and inventory connectors with powerful credentials | New credentials per integration; revoke old ones at decommissioning |
| Push certificates and enrolment tokens | Expiry or ownership under a departed employee's account stops management | Register each with owner and expiry; create new ones under a shared organisational account |
| Audit logs and remote action history | Evidence of wipes, locks and policy changes lost with the old tenant | Export to the SIEM or records store before decommissioning |
Designing the parallel run
Most organisations run both platforms for weeks or months. The design question is how the identity provider treats a device during that window.
- Dual compliance source: configure access policy to accept a compliant state from either platform, for each platform's own enrolled population, for a fixed and dated window. Write the end date into the change record.
- No silent fallback: do not switch device compliance checks to report only to smooth the migration. If a temporary exception is needed for a wave, scope it to named groups and a short expiry, and log every sign-in that uses it.
- Single manager per device: most operating systems allow only one device management enrolment at a time. Make the unenrol and enrol sequence one scripted user journey, and track any device that completes the first step but not the second.
- Old console in maintenance mode: no new policies, no new applications and no new administrators on the old platform once the first wave starts, apart from emergency fixes through change control.
How the compliance signal drives access is covered in the Device Trust and Posture Guide; this playbook only covers keeping it truthful during the overlap.
Choosing a re-enrolment method
| Method | How it works | Strengths | Risks to manage | Best for |
|---|---|---|---|---|
| Wipe and automated re-enrolment | Reassign the device in the maker's enrolment programme, factory reset, and let it enrol to the new server at setup | Clean state; supervision and enrolment restrictions applied from the start; no residue from the old platform | User data loss if not synced; downtime per device; needs every device registered in the programme | Corporate phones and tablets; shared and kiosk devices |
| User-driven unenrol and enrol | User removes the old management profile or agent and enrols with corporate credentials and MFA | No wipe; can be self-service at scale | Devices stuck between platforms; weaker enrolment assurance; support load | Laptops with synced user data; personally owned devices |
| Remote unenrol with agent hand-off | The old platform runs a removal task and installs the new agent or enrolment package | Low user effort for desktops | The old platform's deployment rights are used one last time across the fleet; failed hand-offs leave devices unmanaged | Managed desktop fleets with reliable connectivity |
| Replace at refresh | Enrol new hardware straight into the new platform; retire old devices on schedule | No migration work per device | Long overlap; old platform kept alive for the tail | Fleets close to their refresh date |
The NCSC recommends zero-touch enrolment for organisation-owned devices, and strong authentication for users and devices, including MFA, where administrator or self-enrolment is used. Apply that to the new platform from the first wave: require MFA at enrolment, restrict enrolment to corporate identities, and block personal device types from the corporate-device enrolment profile. For personally owned devices, re-read the BYOD policy before the move: a migration is a chance to switch from full device enrolment to app protection where full management was never needed. See the BYOD Policy Template.
Where identity fits: platform credentials and connectors
A device management platform is wired into the identity estate through non-human identities, and a migration duplicates all of them.
- Identity provider connector: the credential that sends compliance state to the identity provider. Issue a new one for the new platform and remove the old one when the dual compliance window closes, so a stale platform cannot keep vouching for devices.
- Certificate authority connector: the account that requests device and user certificates on behalf of enrolled devices. Scope it to the certificate templates the new platform needs. See the IoT and Device Certificate Provisioning (EST, SCEP) Guide.
- Directory and on-premises connectors: service accounts that join devices to a directory or sync groups. Treat them as privileged service accounts with owners and rotation dates. See the Service Account Security Guide.
- Automation and API clients: scripts and integrations that call the platform API to retire devices, read inventory or open tickets. Inventory every client in the old platform and decide which are rebuilt; do not copy a full-write API key into new scripts for convenience.
- The migration tooling identity: any account used to read the old platform and drive bulk actions should be dedicated, time limited and logged, and removed when the migration ends.
Misuse of a management console or its API to push software or commands across a fleet maps to MITRE ATT&CK T1072 Software Deployment Tools, and theft of integration tokens to T1528 Steal Application Access Token. Both platforms are exposed to this during the overlap.
The phased playbook
- Inventory and baseline. Reconcile devices against the asset register, list every policy, certificate, integration, push certificate, enrolment token and administrator, and record a target setting for each compliance and configuration control.
- Build the new platform securely. Single sign-on and phishing-resistant MFA for all administrators, scoped roles, a separate role for destructive actions, audit logging to the SIEM, and new integration credentials, before any device enrols.
- Rebuild and compare policies. Recreate compliance, configuration and app protection policies from the baseline. Compare each setting with the old platform and document every intentional difference.
- Prove recovery. Enrol test devices, confirm encryption recovery keys and lock bypass codes escrow correctly, and test retrieving a key, a remote lock and a selective wipe.
- Pilot wave. Move IT staff devices first, then a business pilot group covering each operating system and ownership type. Hold a go or no-go review on the gap count, support tickets and compliance accuracy.
- Production waves. Move by site, business unit or device type in waves sized to support capacity. Reassign automated enrolment records per wave, not all at once, so unmigrated devices that are reset do not land on a platform that has no policy for them yet.
- Chase the tail. Report daily on devices still on the old platform and devices on neither. Contact owners, then block access for devices that miss the deadline, rather than extending the dual compliance window indefinitely.
- Close the dual compliance window. Remove the old platform's identity connector so only the new platform can mark a device compliant.
- Decommission the old platform. See the next section.
Rollback: for each wave, define what happens if the new platform fails: devices already moved stay on the new platform with a temporary scoped exception, and remaining waves pause. Keep the old platform's push certificates, enrolment tokens and policies intact until the last wave closes, because re-enrolling devices back is the only true rollback, and it needs a working old platform.
Decommissioning the old platform
An old console that nobody watches can still wipe or lock any device it still manages, and still holds recovery keys and integration credentials.
- Prove it is empty: confirm zero devices report to the old platform, and that every device in the asset register is either in the new platform or recorded as disposed.
- Preserve evidence: export audit logs and remote action history to the SIEM or a records store for the retention period your audit obligations require.
- Handle recovery keys: archive the old recovery key set only for devices not yet confirmed in the new escrow, under restricted access, and destroy it once confirmed.
- Revoke everything it trusted: integration credentials, API clients, connector service accounts, certificates issued through it, and administrator accounts.
- Release its enrolment assets: remove the old server from the device maker's enrolment programme and revoke its push certificates and enrolment tokens, so no device can enrol to it again.
- Close the contract cleanly: for a hosted platform, confirm data return and deletion with the provider. NIST CSF 2.0 outcome GV.SC-10 expects supply chain risk plans to cover activities after a partnership ends. For on-premises servers, sanitise storage in line with NIST SP 800-88 Rev. 2, Guidelines for Media Sanitization.
Decision guidance
| Choice | Option A | Option B | Choose A when |
|---|---|---|---|
| Cutover style | Big bang over a weekend | Waves over weeks | Fewer than a few hundred devices, one site and simple policies; most organisations need B |
| Corporate mobile devices | Wipe and automated re-enrolment | User-driven unenrol and enrol | Devices are in the maker's enrolment programme and data is synced to corporate cloud services |
| Personally owned devices | Re-enrol with full device management | Move to app protection only | Regulation or data classification genuinely requires device-level control |
| Policy approach | Like-for-like copy | Redesign against a current baseline | Policies were reviewed against a benchmark in the last year |
| Overlap length | Short, with a hard deadline and access blocking | Long, until hardware refresh | Support capacity allows it; a long overlap doubles the attack surface |
Small organisations: the biggest wins are confirming recovery keys before unenrolment, moving push certificates and enrolment tokens to an organisational account, and switching off the old console on a fixed date. Large organisations: expect several operating systems, regional teams and dozens of integrations; maintain a migration register per wave with owners, and give the security team a daily view of the gap count.
Detection and measurement
Forward audit logs from both platforms to the SIEM for the whole migration. Alert on remote wipe, lock or retire actions on the old platform after it enters maintenance mode, bulk actions outside approved waves, new administrators or API clients on either platform, policy changes that weaken compliance settings, use of the old platform's integration credentials after the dual compliance window closes, and sign-ins that relied on a temporary device exception.
- Unmanaged gap: devices in the asset register reporting to neither platform, counted daily (target zero beyond devices actively in a scheduled wave).
- Recovery key coverage: percentage of migrated devices whose encryption recovery key is confirmed in the new escrow (target 100 percent before the old key set is destroyed).
- Policy parity: percentage of baseline settings enforced at equal or stronger values in the new platform.
- Compliance accuracy: migrated devices evaluated compliant within 24 hours of enrolment, as a share of devices enrolled in the period.
- Exception use: sign-ins using a temporary migration exception, trending to zero by the published end date.
- Residual trust: credentials, connectors and administrators remaining on the old platform at decommissioning (target zero).
After the migration, the recurring measures in the MDM and UEM platform health check take over.
Common failure patterns
- Report-only compliance left on: device checks are relaxed to avoid lockouts during the move and never turned back on.
- Recovery keys stranded: devices unenrol before their encryption recovery keys are escrowed in the new platform, then the old tenant is deleted.
- Devices in limbo: users complete the unenrol step but not the enrol step, and those devices keep accessing email through cached tokens.
- Automated enrolment reassigned too early: the whole fleet is pointed at the new server before policies exist, so any reset device enrols with no restrictions.
- Old console forgotten: the previous platform is left running with former administrators and integration credentials for months after the last device moved.
- Tokens owned by a person: push certificates or enrolment tokens are created under an individual's account and expire after that person leaves.
Framework mapping
| Migration activity | NIST SP 800-53 Rev. 5 | NIST CSF 2.0 | ATT&CK |
|---|---|---|---|
| Device inventory and reconciliation | CM-8 System Component Inventory | ID.AM-01 | |
| Policy rebuild and parity | CM-6 Configuration Settings; AC-19 Access Control for Mobile Devices; CM-3 Configuration Change Control | PR.PS-01 | |
| Recovery key escrow | SC-28 Protection of Information at Rest | PR.PS-01 | |
| Administrators and destructive actions | AC-2 Account Management; AC-6 Least Privilege | PR.AA-05 | T1072 Software Deployment Tools |
| Connectors and API credentials | IA-5 Authenticator Management | PR.AA-01 | T1528 Steal Application Access Token |
| Logging across both platforms | AU-2 Event Logging; AU-6 Audit Record Review, Analysis, and Reporting; AU-11 Audit Record Retention | PR.PS-04, DE.CM-09 | |
| Supplier exit and server disposal | SA-9 External System Services; MP-6 Media Sanitization | GV.SC-10 |
In CIS Controls v8.1 the work sits mainly in Control 1 (Inventory and Control of Enterprise Assets) and Control 4 (Secure Configuration of Enterprise Assets and Software), with administrator and connector accounts under Control 5 (Account Management) and Control 6 (Access Control Management).
Practitioner checklist
- Reconcile the device inventory against purchase records and the asset register before the first wave.
- Build the new platform with SSO, phishing-resistant MFA, scoped roles and SIEM logging before any device enrols.
- Rebuild policies from a documented baseline and record every setting difference.
- Confirm encryption recovery key escrow and test a key retrieval, lock and selective wipe on pilot devices.
- Configure a dated dual compliance window; never switch device checks to report only.
- Reassign automated enrolment records per wave, after the target policies exist.
- Issue new credentials for every connector and API client, and move push certificates and enrolment tokens to an organisational account.
- Freeze the old console: maintenance mode, minimal named administrators, alerts on any remote action.
- Track the unmanaged gap daily and block devices that miss the deadline.
- Decommission deliberately: export logs, revoke credentials and enrolment assets, and confirm provider deletion or sanitise servers.
Standards and references
- NIST: SP 800-124 Rev. 2, Guidelines for Managing the Security of Mobile Devices in the Enterprise
- NCSC: Device security guidance, Mobile Device Management
- NCSC: Device security guidance, Erasing devices
- NIST: SP 800-88 Rev. 2, Guidelines for Media Sanitization
- NIST: SP 800-53 Rev. 5, Security and Privacy Controls for Information Systems and Organizations
- NIST: Cybersecurity Framework 2.0
- CIS: Critical Security Controls Version 8.1
- MITRE ATT&CK: Software Deployment Tools (T1072)
Related NHI Mgmt Group resources: Mobile Device Management Security Guide · If your MDM and UEM platform is compromised · UEM Buyer's Guide · Device Reassignment and Reimaging