Microsoft keeps your client’s tenant running. Keeping their data recoverable is your job. Exchange, OneDrive, SharePoint, Teams and Entra ID, per protected user, storage included, retention infinite while the subscription is active.
Native retention windows are measured in days. A compromised admin can wipe a mailbox, and Microsoft’s policies protect the tenant. Your client’s four years of SharePoint are your problem, and they will not remember whose responsibility it technically was. When that call comes, it comes to you.
Retention that expires quietly
Native recycle bins hold items for days. The request arrives later.
Backup tiers with asterisks
Unlimited with a fair-use multiplier has a footnote. You find it at renewal.
Departed users, unpriced
An offboarded employee is a licence you keep paying for, or data you lose.
Infinite while active
Rolling immutable locks renew daily for as long as protection is active and paid. Fixed presets exist only where compliance requires deletion, never as a price lever. The locks survive a compromised console: a stolen admin session cannot shorten them.
Archived users, priced in the open
Departing employees become a named, billable, searchable state with their own terms and legal holds, never an unpriced ambiguity your renewal discovers later.
From one item to the whole tenant
Roll a tenant, a department, selected users or one workload back to a single cross-workload point. Same restore grammar as everything else.
Restores that end in evidence
Every restore closes with a signed manifest: what was recovered, verified against what, signed by whom. Your client’s auditor gets a document they can file.
What a Teams restore can and cannot do
Microsoft provides no API to rewrite channel history. Restored messages land in a migration channel, and every run produces a complete evidence export.
Search across everything
Mail, files, sites and Teams in one index. Preview an item, restore it, or export it, and filter to what changed during an incident.
Every workload family carries a stated classification. Where Microsoft has no faithful write API, this table says so.
| Workload | What is captured | How it comes back |
|---|---|---|
| Exchange Online |
|
|
| OneDrive for Business |
|
|
| SharePoint Online |
|
|
| Microsoft Teams |
|
|
| Microsoft Entra ID |
|
|
One queue for every customer, with the evidence attached to the alert that explains it.
| Mode | When you reach for it | What it protects you from |
|---|---|---|
| Browse and search | You know roughly what was lost but not exactly where | Restoring more than you meant to |
| Non-destructive restore | The user still needs what is there now | Overwriting live work |
| Original location | The item should return exactly where it was | Confusing the user with a new location |
| Alternate location | The original mailbox, site, or user is gone | Being blocked by a deleted destination |
| Signed export | Legal needs evidence, or the tenant is unavailable | Depending on Microsoft to read your own data |
| Tenant-scale rollback | An attack or a bad change crossed many users | Rebuilding a tenant by hand |
One GDAP partner consent covers every tenant your relationships already cover. No per-tenant ceremony, no credential chase.
Accept the recommended policy. Every workload starts capturing and the immutability locks begin renewing daily.
Restore an item in front of the client and hand them the signed manifest.
You hand them a signed manifest. Connect with one GDAP consent and every restore ends in evidence. The question becomes a file you send back.
One GDAP consent, the recommended policy, and a restore you can demonstrate.
