Lost Device → Remote Lock
Lock issued within 5 minutes of report
A laptop is missing. Reported at 6pm on a Friday. The device holds an active SSO session, cached credentials, potentially unencrypted files. Every hour until it's locked is an hour that data is exposed. IT support isn't available till Monday. The employee doesn't know what to do. This is where a small process failure becomes a real security incident.
An hour-by-hour walkthrough.
Step by step.
- 01
Verify identity of caller
Read requester identity from Slack / Teams SSO. Step-up MFA on requester's registered phone (not the missing laptop). Verify device ownership in Context Graph.
Slack · Okta · Context Graph - 02
Locate device
Look up device serial in Jamf / Intune / Kandji. Retrieve last check-in location if MDM has geolocation enabled. Confirm device is still in the requester's assignment.
Jamf · Intune · Kandji - 03
Issue remote lock
MDM API call to lock the device immediately. Active SSO sessions invalidated. Screen locked with a message including recovery contact info for a would-be Good Samaritan.
Jamf · Intune · Kandji · Okta - 04
Confirm status + open incident
Confirm lock succeeded. Open security incident ticket with device metadata + timeline. Page security on-call for after-hours reports. Automatically offer loaner device via RMA.
Slack · Security queue · RMA workflow - 05
Follow-up (recovery or wipe)
72-hour countdown until auto-wipe. If device is recovered, requester confirms, Ivy cancels the wipe with a re-verification. If not, wipe fires, RMA replacement ships, and audit closes the incident.
MDM · Slack · RMA workflow
What you connect to make this run.
Jamf Pro · Intune · Kandji
read+writeMDM API credential with lock + wipe + geolocation scopes. Lock is immediate; wipe is time-gated with human override.
Okta · Microsoft Entra ID
writeSession-invalidation on the device. Even if the finder cracks the OS lock, SSO won't renew — attacker gets nothing durable.
Security queue (PagerDuty · Incident.io)
writeAfter-hours reports page security on-call. Weekday reports create ticket + notify security via Slack. Every report is treated as a potential incident until resolved.
RMA workflow
writeAutomatic loaner-device offer + replacement RMA. Prevents "lost laptop delays productivity for a week" pattern.
Before and after, honestly.
Playbooks that pair with this one.
Device Recovery / Unlock
The inverse — user has device but can't get in.
RMA & Repair
The replacement path when the device doesn't come back.
Compromised Credential Response
Fires if any credentials were exposed on the device.
Anomalous Login
Watches for login attempts on the locked device or accounts.
Answers about this playbook.
What if the device is BYO (personal-owned, MDM-enrolled)?
Same flow with BYO-appropriate limitations. Ivy can lock and wipe the corporate profile; the personal side of the device is untouched. Employee retains their personal data + apps.
How does this handle a rushed "actually I found it" report?
Unlock requires identity re-verification. Requester DMs Ivy, Ivy re-verifies via step-up MFA, cancels the wipe countdown, delivers the FileVault recovery key. Full audit trail of every lock and unlock.
What about compliance reporting?
Every lost-device report generates a compliance record: device metadata, incident timeline, resolution (recovered / wiped / replaced), any data exposure. Ready for SOC 2 / ISO 27001 evidence.
Can the device geolocation be used to actually find the device?
For "lost on the train" cases: unlikely to help unless MDM ping happens while the device is on. For "stolen" cases: geolocation data included in the police report (some MDMs support 'lost mode' with periodic geolocation ping to a covert URL).
What if the report is fraudulent (someone else reporting)?
Identity verification is designed to catch this. The requester must be the device's assigned owner in Context Graph and pass step-up MFA. If either fails, Ivy declines and escalates the mismatched request to security for review.
See it run on your data.
Free plan, no credit card. Connect the systems this playbook needs and run it against a past event first.