Test a pilot before restoring a larger batch

Choose a useful sample

• A file shared directly with an internal user for Read.
• A different file shared for Edit, plus a folder whose children inherit access.
• A child item with unique permissions if your migration includes them.
• A guest recipient and a group share if those are part of the approved scope.
• Each sharing-link type you plan to recover, including an expiring example when applicable.

Before approval

1. Open the destination files and confirm they are the intended copies.
2. Record the source path, destination path, expected recipient, and expected access level.
3. Check the connected tenant IDs and the Restore shares to groups choice.
4. Review every proposed pilot change. Resolve unknown identities and confirm external-guest decisions.
5. Record credits available and the estimated credits needed. Remember that permission rows are not the credit count.

After execution

1. Check Verify results and open any failed or unverified detail.
2. Sign in as representative destination recipients in a separate browser session. Confirm they can open the expected content.
3. Test Read and Edit behavior separately. A recipient who already has broader access through another group may not demonstrate the restored permission by itself.
4. For links, use the destination link from the link map. Confirm its audience and expiration where applicable.
5. Confirm successful already-restored items do not consume another credit when safely retried or included in a later run.
6. Save the results and any unresolved exceptions before expanding the scope.

When to pause the rollout

Pause if a destination path is wrong, an unexpected recipient would gain access, a role differs from the intended source access, or a result cannot be verified. Fix the matching or policy issue and prepare a fresh plan. More rows with the same problem will not make the pilot conclusive.