schedule 15-min read

Microsoft 365 Direct Send, Printers and DMARC: How to Fix Scan-to-Email

DMARC AI editorial team · Last updated

Why scanners and apps fail DMARC in Microsoft 365, and how to pick SMTP AUTH, Direct Send or a relay connector before moving clients to p=reject.

01

Why do printers and scanners fail DMARC in Microsoft 365?

Printers, scanners and line-of-business apps fail DMARC in Microsoft 365 when they send as the client's domain without authenticating. With Direct Send, the device connects to the MX endpoint like any internet sender, so its IP must be in SPF and nothing signs DKIM. Once DMARC reaches p=quarantine or p=reject, Microsoft treats that mail as self-to-self spoofing.

This guide covers the sending methods Microsoft documents for devices and applications, what each one does to SPF (Sender Policy Framework), DKIM (DomainKeys Identified Mail) and DMARC (Domain-based Message Authentication, Reporting and Conformance), the Reject Direct Send setting, and the SMTP AUTH Basic authentication timeline. It ends with a test procedure and an MSP (managed service provider) checklist. The full tenant setup is in the DMARC for Microsoft 365 implementation guide; protocol background is in the DMARC guide.

02

Key takeaways

  • Microsoft documents three main methods for devices and apps: SMTP AUTH client submission, Direct Send and SMTP relay through a connector. Only SMTP AUTH and the relay can reach external recipients.
  • Direct Send is unauthenticated. Its DMARC result depends entirely on the device's public IP being in SPF, and Microsoft says "most customers don't need to use Direct Send".
  • Set-OrganizationConfig -RejectDirectSend $true blocks anonymous mail from your own accepted domains that matches no inbound connector. Microsoft still describes it as a Public Preview.
  • Basic authentication for SMTP AUTH is due to be disabled by default for existing tenants at the end of December 2026, with admins able to turn it back on.
03

Which sending options does Microsoft document for devices and apps?

Microsoft's multifunction device and application guide (last updated 16 September 2026) describes three classic options plus High Volume Email (HVE). They differ in endpoint, authentication, licensing and whether mail can leave the tenant.

SMTP AUTH client submission Direct Send SMTP relay (connector)
Endpoint smtp.office365.com MX endpoint, e.g. example-co-uk.mail.protection.outlook.com MX endpoint
Port 587 (recommended) or 25 25 25
TLS Required (TLS 1.2 or 1.3) Optional Optional in the comparison table
Authentication Mailbox credentials, OAuth recommended None Inbound connector: TLS certificate (recommended) or static public IP
Licensed mailbox Yes No, but the From address must use an accepted domain No
External recipients Yes No: "Mail sent to recipients outside your cloud-based organization is rejected" Yes
Limits 10,000 recipients per day, 30 messages per minute Standard anonymous throttling "Reasonable limits", no spam or bulk mail
Spam filtering Bypasses most spam checks for internal recipients All scanning applies, like anonymous internet mail Not bypassed; Microsoft recommends an SPF entry

HVE is a fourth method for large internal volumes on port 587. It cannot relay to the internet, so it does not help a scanner that emails customers. Microsoft does not support port 465 for client submission, and says the IIS SMTP server in Windows Server is "unsupported for relaying email" to Microsoft 365.

04

What happens to SPF, DKIM and DMARC with each option?

The outcome depends on who connects to the receiving server and who signs. With SMTP AUTH and the relay, Exchange Online sends the mail out, so Microsoft's IPs and Microsoft's DKIM signing apply. With Direct Send, the device itself is the sender Microsoft evaluates.

Device or app Option Auth SPF / DKIM outcome DMARC alignment Pros Cons
Modern MFP with TLS 1.2 and OAuth SMTP AUTH Mailbox, OAuth SPF passes via include:spf.protection.outlook.com; DKIM d= client domain if enabled Aligned SPF and DKIM Internal and external, Sent Items Licensed mailbox; Basic auth timeline
Older MFP, no modern TLS Direct Send None SPF passes only if the site's static IP is in SPF; no DKIM SPF only, and only if the IP is listed No licence, no credentials Internal only; blocked by Reject Direct Send
Fleet of scanners or an on-premises app server SMTP relay Connector, certificate or static IP External: SPF via Microsoft's include, DKIM by Exchange Online if enabled Aligned when the envelope and From use the client domain External delivery, no mailbox Needs a static IP or certificate; not for cloud-hosted apps
SaaS or app hosted in a cloud such as Azure SMTP AUTH (OAuth) or the vendor's own DKIM Mailbox or vendor Depends on the path Aligned only with vendor DKIM or SMTP AUTH Avoids relaying through the tenant Relay connectors cannot be used from third-party hosted services

SMTP AUTH client submission

The device signs in as a mailbox, so Exchange Online treats the message like mail from that user. The envelope sender (MAIL FROM) is the mailbox address, so SPF passes through Microsoft's include and aligns with the From domain. If DKIM is enabled for the domain, the message carries a d= signature for the client's domain.

The licensed mailbox needs SMTP AUTH enabled. Microsoft disabled SMTP AUTH for organisations created after January 2020, and it is off when Security defaults are on. If the device sends as a different address, the sign-in account needs Send As permission, or it gets 5.7.60 SMTP; Client doesn't have permissions to send as this sender.

Direct Send

Direct Send is anonymous SMTP to the tenant's MX endpoint. Microsoft says these messages "are just like anonymous email from the internet (all scanning and protections apply)". SPF is checked against the device's public IP, so a dynamic IP or a shared office NAT address fails. Nothing adds DKIM. Microsoft recommends Direct Send "only for advanced customers willing to take on the responsibilities of email server admins".

SMTP relay through a connector

The device sends to the MX endpoint, and an inbound connector recognises it by certificate or static IP. The connector attributes the mail to the organisation, which lets it relay to external recipients. Outbound, receivers see Microsoft's IPs and Exchange Online's DKIM signature. Microsoft still tells you to add the device's IP to SPF, so inbound filtering does not mark the mail as spam.

05

Why is internal Direct Send mail treated as spoofing?

Because the From domain is one of the tenant's own accepted domains and the message arrived unauthenticated. Microsoft calls this intra-org or self-to-self spoofing. If SPF does not pass for the device IP, DMARC fails, and the result depends on the client's published policy.

Microsoft's anti-spam message headers page defines the composite authentication (compauth) reasons you will see:

Reason Microsoft's meaning What it usually means for a scanner
010 Failed DMARC, policy is p=reject or p=quarantine, and the sending domain is an accepted domain (self-to-self) The client went to enforcement and the device is not in SPF
601 Failed implicit authentication; the sending domain is an accepted domain Still at p=none, device unauthenticated
1xx Passed explicit or implicit authentication Device IP is in SPF, or it uses SMTP AUTH

The anti-spoofing protection page adds that intra-org spoofing shows CAT:SPOOF and SFTY:9.11 in X-Forefront-Antispam-Report. At p=none, spoof intelligence decides whether to deliver it. At enforcement, the anti-phishing policy applies its DMARC action, which is why "scan to email stopped working" tickets arrive the day a client moves to p=reject. Allowing the domain pair in the Tenant Allow/Block List is a short-term workaround, not a fix. DMARC in Microsoft Defender explains those policy actions.

06

What happens to DMARC for mail that goes outside the tenant?

Direct Send cannot reach external recipients, so external DMARC only applies to SMTP AUTH and relay mail. For both, Exchange Online is the sending server. SPF passes when the envelope domain is the client's domain and the record includes spf.protection.outlook.com. DKIM passes when signing is enabled for the From domain.

Under RFC 9989, DMARC passes if either SPF or DKIM passes for a domain that aligns with the From domain. A relay device whose envelope address is outside the client's domains loses SPF alignment, so DMARC then depends on DKIM alone. The DKIM configuration page says Microsoft 365 does not sign outbound mail from custom domains until you enable it, and omits the signature when sender and recipient are in the same organisation. Test DKIM with an external mailbox.

07

What should go in SPF for devices and apps?

Add only the public IP that connects to Microsoft's MX endpoint, and only for Direct Send or IP-based relay. SMTP AUTH needs nothing extra, because Exchange Online sends the mail and the existing include covers it. Use ip4, which costs no DNS lookups under RFC 7208.

example.co.uk.  3600  IN  TXT  "v=spf1 ip4:203.0.113.25 include:spf.protection.outlook.com -all"

Microsoft's troubleshooting guide for printers, scanners and LOB apps warns that "a network or ISP change could change your static IP address", and that you must update SPF when it does. Record the IP against the site and the ISP circuit in your documentation. The rest of the record, including lookup counts, is covered in the Microsoft 365 SPF record explained. Do not add an include for a cloud LOB vendor unless that vendor uses your domain in the envelope; aligned vendor DKIM is usually the better fix, as handling third-party senders explains.

08

How do I configure each option?

Start with SMTP AUTH where the device supports TLS 1.2 and, ideally, OAuth. Use a certificate-based relay connector for fleets or devices that need external delivery without a mailbox. Keep Direct Send for legacy internal-only devices you cannot replace yet.

SMTP AUTH: enable it for one dedicated mailbox only. These commands come from Microsoft's Enable or disable SMTP AUTH page. The mailbox setting overrides the organisation setting.

# Keep SMTP AUTH off organisation-wide
Set-TransportConfig -SmtpClientAuthenticationDisabled $true

# Enable it for the scanner mailbox only
Set-CASMailbox -Identity scanner-hq -SmtpClientAuthenticationDisabled $false
Get-CASMailbox -Identity scanner-hq | Format-List SmtpClientAuthenticationDisabled

Device settings: server smtp.office365.com, port 587, STARTTLS on, the mailbox's credentials or OAuth token.

SMTP relay: create the connector. Microsoft documents the wizard in the Exchange admin center under Mail flow > Connectors: from "Your organization's email server" to Office 365, authenticated by certificate subject (Microsoft suggests a value such as *.contoso.com) or by a static IP that belongs only to you. A PowerShell sketch of the IP-based version, to adapt and test:

New-InboundConnector -Name "HQ scanner relay" -ConnectorType OnPremises `
  -SenderDomains * -SenderIPAddresses 203.0.113.25 -RestrictDomainsToIPAddresses $true
Get-InboundConnector | Format-List Name,ConnectorType,SenderIPAddresses,TlsSenderCertificateName

Point the device at the tenant's MX host on port 25. Microsoft does not allow dynamic IPs, and says not to share the IP "with anyone outside of your organization". If a gateway sits in front of Microsoft 365, the device may be relaying through the gateway instead; check Enhanced Filtering for Connectors and the gateway article before changing routing.

Direct Send: if you must keep it. Point the device at the MX host on port 25, use a static IP, add that IP to SPF and confirm the device sends with a client address in both envelope and From.

09

What is Reject Direct Send and should I enable it?

Reject Direct Send is an Exchange Online organisation setting that refuses anonymous mail claiming to be from your own domains. Per the Set-OrganizationConfig reference, with $true Exchange Online rejects anonymous messages "sent from your own domain to your organization's mailboxes" when two conditions are met:

  • the message matches no inbound connector configured to match the sender IP or certificate;
  • the domain in the MAIL FROM (P1, envelope sender) address is an accepted domain.

Get-OrganizationConfig | Format-List RejectDirectSend
Set-OrganizationConfig -RejectDirectSend $true

Microsoft's Exchange Team post gives the rejection as 550 5.7.68 TenantInboundAttribution; Direct Send not allowed for this organization from unauthorized sources. In its latest revision (20 May 2026) the post still calls the feature a Public Preview with no fixed general availability date, says it is off by default, and says Microsoft plans to turn it on by default for new tenants. Subdomains of an accepted domain are affected when that domain has subdomain matching enabled. The multifunction guide also says Microsoft is "working on an option to disable Direct Send by default". Check the Microsoft 365 Message Center before you rely on these statuses.

Two cautions. The check uses the envelope sender, not the visible From header, so it does not replace DMARC enforcement. And every legitimate Direct Send device in the tenant breaks the moment you enable it. Move those devices to SMTP AUTH or a connector first. Microsoft's post mentions a Change Optics Report in the Exchange admin center (also in preview) for seeing Direct Send traffic that would be affected.

10

What is the status of Basic authentication for SMTP AUTH?

As of 15 November 2026, Microsoft's published plan is to disable Basic authentication for SMTP AUTH by default in existing tenants at the end of December 2026. Admins will still be able to turn it back on. New tenants created after that will not have it by default. Microsoft says it will announce the final removal date in the second half of 2027.

That timeline comes from the Exchange Team's updated SMTP AUTH deprecation timeline, published 27 January 2026 and revised 29 January 2026. It replaced an earlier March 2026 date. We found no newer official change, but this timeline has moved before, so recheck it before planning. Microsoft's Basic authentication deprecation page points to the same post.

For MSPs, the practical point is simple. Any scanner using SMTP AUTH with a username and password is likely to fail in January 2027 unless someone re-enables Basic auth. Use the remaining weeks to update firmware for OAuth, or move the device to a relay connector.

11

How do I test a device before enforcement?

Send two test messages from each device, one to an internal mailbox and one to an external mailbox you control. Then read Authentication-Results on both. The internal copy shows how Microsoft classifies the device; the external copy shows SPF, DKIM and DMARC as other receivers will see them.

  1. Send the tests with a clear subject such as DMARC test HQ-MFP-02.
  2. Trace the message in the Exchange admin center or with Get-MessageTraceV2, and note the connector the message was attributed to.
  3. Read the headers. Paste them into the email header analyzer, or follow how to read email headers.
  4. Compare with the expected result for the chosen option.

A Direct Send device missing from SPF, after the client moves to p=reject, looks like this internally:

Authentication-Results: spf=fail (sender IP is 203.0.113.25)
 smtp.mailfrom=example.co.uk; dkim=none (message not signed)
 header.d=none;dmarc=fail action=oreject header.from=example.co.uk;
 compauth=fail reason=010

The same device moved to a relay connector, received at an external mailbox, should show spf=pass for example.co.uk, dkim=pass with header.d=example.co.uk and dmarc=pass. Check the SPF record itself with the SPF analyzer.

12

What should an MSP check before moving a client to p=quarantine or p=reject?

Treat devices as a named workstream in the enforcement plan, not an afterthought. They send small volumes, mostly internally, so they barely show in aggregate reports. The client discovery guide explains how to find them; this checklist covers what to do once you have.

  1. List every device and app that sends as the client: MFPs, scanners, alarm panels, backup and monitoring alerts, ERP and LOB servers.
  2. Record the method per device: SMTP AUTH, Direct Send or connector, with endpoint, port, IP and mailbox.
  3. Check TLS and OAuth support in firmware for every SMTP AUTH device, against the December 2026 Basic auth change.
  4. Move Direct Send devices to SMTP AUTH or a connector wherever they need external delivery or can be changed.
  5. Fix SPF for any remaining Direct Send or IP-based relay sources, using static IPs only.
  6. Enable DKIM for every sending custom domain and confirm header.d with an external test.
  7. Test every device with the header procedure above and keep the result in the ticket.
  8. Decide on Reject Direct Send once no legitimate Direct Send traffic remains, and record the change.
  9. Move the policy one step at a time: p=none, then p=quarantine, then p=reject, optionally rehearsing each step with t=y from RFC 9989.
  10. Watch the first week for intra-org spoof detections and new failing sources.

The wider safety checks are in how MSPs can avoid breaking client email with DMARC, and the policy move itself is in moving from monitoring to enforcement safely. For UK clients, the NCSC's email security and anti-spoofing guidance treats SPF, DKIM and DMARC as one set of controls, which helps when explaining device changes to a client.

13

FAQ

Can Direct Send deliver to Gmail or other external recipients?

No. Microsoft states that Direct Send "sends email to Microsoft 365 or Office 365 recipients only" and that mail to recipients outside the organisation is rejected. If a scanner or application must email customers or suppliers, use SMTP AUTH client submission or an SMTP relay connector instead.

Does Reject Direct Send replace DMARC enforcement?

No. Reject Direct Send checks the envelope sender domain and only covers anonymous mail that matches no inbound connector. Mail with a different envelope domain, or mail arriving through a connector, is not covered. DMARC checks the visible From domain at every receiver, including receivers outside Microsoft 365, so you still need p=reject.

Do I need to add smtp.office365.com to SPF for scanners?

No. Mail submitted through SMTP AUTH leaves through Exchange Online, so include:spf.protection.outlook.com already covers it. Add an ip4 entry only for a device or server that connects directly to the MX endpoint, which means Direct Send or an IP-based relay connector, and only for a static public IP.

Will Microsoft DKIM-sign mail from a scanner?

Yes, for SMTP AUTH and relay mail that Exchange Online sends out, once DKIM signing is enabled for the domain in the From address. Direct Send mail is never signed, because it arrives as inbound internet mail. Microsoft also omits the signature between senders and recipients in the same organisation.

14

Next steps

Inventory every device and app per client, move them off Direct Send and password-based SMTP AUTH where you can, and test each one before the next policy step. DMARC aggregate reports show which sources still fail SPF or alignment; DMARC AI processes those reports across client tenants and highlights failing sources and alignment problems, so remaining devices stand out before p=reject.

Related articles

Related tools

Ready to Implement?

Get authenticated mail moving in minutes: start free, book a guided demo, or talk to the team about your stack.