[ Enterprise operations ]

Cloud email migration: inventory the systems outside the tenant

Inventory application mail, relays and delegated senders before moving cloud mailboxes so the migration covers the entire mail estate.

By the EmailAuth research team · 3 min read

[ Key points ]

  • Inventory application mail, relays and delegated senders before moving cloud mailboxes so the migration covers the entire mail estate.
  • A completed mailbox cutover does not prove application migration; verify destination scope and every remaining relay dependency.
On this page

A cloud mailbox migration can finish successfully while invoices, alerts and scanned documents stop arriving. Those workflows may send through devices, applications or third-party providers outside the mailbox tenant. Build the migration inventory around business mail paths, not only the list of user accounts being moved.

Map the systems the tenant list misses

Ask each business and infrastructure team which systems send messages. Include customer communications, billing, identity workflows, incident alerts, printers, website forms, scheduled reports and SaaS campaigns. Record purpose, owner, criticality, sending method, network dependency and evidence of the last end-to-end test.

For each path, distinguish visible From, envelope sender, DKIM signing domain, destination scope and bounce handling. Document whether the service talks directly to an external provider, an on-premises relay or a tenant endpoint. Microsoft explicitly documents several application-mail methods with different prerequisites and restrictions. Direct Send is limited to recipients in the organization; SMTP relay and authenticated submission have different setup requirements. Microsoft guidance. Verify the current supported method for the actual provider rather than copying another platform's settings.

Connect transport to authentication

Changing the mailbox platform does not automatically update an application's sender configuration or DNS. Establish who signs each stream, which envelope identity SPF evaluates and how an aligned identifier relates to the visible From domain. RFC 9989. Inventory DKIM selectors and third-party dependencies before removing old records.

Also identify IP-based relay restrictions, certificates, firewall rules, connector dependencies and credential references without exporting the secrets. Record which team owns each dependency. An MX change principally changes inbound routing; it is not a complete migration plan for outbound application mail.

Make recipient scope a migration gate

Suppose customer receipts use a SaaS signing route, a printer uses an internal relay, and an incident application submits through the old platform. The receipt route may remain outside mailbox cutover but still depends on its DNS and provider account. The printer and incident application require their own destination, network and submission checks; migrating employee accounts does not resolve those dependencies.

If the incident application must notify an external on-call address, Microsoft Direct Send is not an appropriate route for that required external delivery. Verify a supported method with the actual account and network prerequisites instead. This hypothetical scope decision does not configure a tenant or test provider eligibility. Preserve a hold until the selected route has authorized end-to-end evidence.

Plan coexistence deliberately

For each service, specify the old path, new path, cutover owner, test cases, failure signals and rollback action. Keep only the parallel authorizations required for the approved coexistence period. Duplicate sends are a separate risk: migrating an application job and leaving its old scheduler active can send the same notification twice even when both messages authenticate.

Exercise internal and external destinations where the business requires them. Test bounce processing, retry behavior and monitoring alerts as well as receipt of one successful message. Retain receiver evidence for signature and alignment results. An SMTP acceptance response alone does not prove the destination mailbox displayed the message or that the application processed a failure correctly.

Set an exit condition for the old platform

Do not decommission the old relay because all user mailboxes moved. Close every dependency row with observed success, an approved exception or an explicit retirement decision. Include scheduled flows that have not run during the cutover window. Export the evidence needed for future troubleshooting before access to the old platform expires.

When retiring permissions, use exact record changes and service configuration changes under normal approval. Remove old credentials and authorizations only after the corresponding service is migrated or retired. State remaining unknowns in the migration decision so incomplete evidence is not mistaken for completion.

[ Takeaway ]

A completed mailbox cutover does not prove application migration; verify destination scope and every remaining relay dependency.

[ By role ]

What this means for you

Choose your role for practical next steps.

IT / Mail Admin

Inventory every path

Track relay, provider and DNS dependencies.

Next steps

  1. Map application endpoints.
  2. Test bounce handling.
  3. Retain old-path evidence.

CISO / SecOps

Limit coexistence exposure

Temporary parallel access needs a clear end.

Next steps

  1. Review permissions.
  2. Track credential retirement.
  3. Document exceptions.

Founder / Non-technical

Scope the whole migration

Critical messages exist outside employee mailboxes.

Next steps

  1. Assign business flow owners.
  2. Confirm external recipient needs.
  3. Approve exit criteria.

[ FAQs ]

Frequently asked questions

Does moving all mailboxes finish migration?

No. Application and device mail may use independent paths.

Does an MX change redirect every outbound application?

No. Applications may retain separate configured endpoints and providers.

Can Direct Send reach external recipients?

Microsoft documents it for recipients in the organization, so it should not be selected for required external delivery.

Should the inventory contain credentials?

No. Record protected-store references and responsible owners instead.

Why test bounces?

A successful message does not establish that failure notifications and retry handling still work.

Can two valid paths create duplicates?

Yes. Coexistence requires clear scheduler and cutover ownership.

Does the example test a live tenant?

No. It reasons from hypothetical service needs and documented recipient scope; configuration and end-to-end tests remain required.

When can the old relay retire?

After every dependent flow has verified migration or an approved retirement decision.

[ Sources ]

Sources

  1. Microsoft application and device mail setup

    Sections Introductory method comparison table · Client SMTP submission: Authenticate mail sent from your device or application through a cloud mailbox · SMTP relay: Configure a connector to relay email from your device or application through Microsoft 365 or Office 365 · Direct Send: Send mail directly from your device or application to Microsoft 365 or Office 365

  2. RFC 9989 DMARC

    Sections §4.4 · §4.10.1 · §5.3.5 · §5.4 · §7.4 · §8

  3. RFC 6376 DKIM

    Sections §3.5 · §3.6 · §5.2 · §6.1.2

  4. RFC 5321 Simple Mail Transfer Protocol

    Sections §4.2.5 · §4.5.3.2.6 · §6.1

[ Glossary ]

Glossary

Coexistence
A controlled interval during which old and new mail paths remain available.
Relay
A server that accepts and forwards mail under configured rules.

Enterprise provenance

Why top financial institutions and governments retired their email security dashboards.

Traditional email authentication platforms sell you a dashboard and ask your engineers to do the labour. In high-stakes environments, manual oversight means delayed enforcement, configuration drift and unmonitored shadow domains.

EmailAuth is the agentic email provenance platform. Our agents enforce p=reject, keep third-party senders aligned and move your domains to full RFC 9989 compliance, the new DMARC standard,, with human-in-the-loop auditability.

Governments Large enterprises Tier-1 financial institutions

SaaS · Dedicated tenancy · Air-gapped on-prem · BYOK encryption

Knowledge Base