How to Triage User-Reported Phishing Emails in Microsoft 365: A Practical SOC Runbook
Investigate user-reported phishing emails in Microsoft 365 with a practical seven-step SOC runbook covering threat analysis, incident prioritization, campaign scoping, containment, and response.
An employee reports an Outlook message from a familiar supplier requesting updated payment instructions. Microsoft 365 delivered it, but the employee senses something is wrong.
What should the security team do next?
An investigation must establish the email's intent, whether anyone interacted, whether other mailboxes are exposed, and whether containment is needed.
This guide provides a repeatable phishing-triage workflow for Microsoft 365 security operations centers (SOCs). It covers employee reporting, Defender evidence, prioritization, campaign scoping, remediation, user feedback, and measurements that help teams improve. Adapt it to your licenses, permissions, and response policies.
What Is Phishing Triage in Microsoft 365?
Phishing triage is the process of receiving a suspicious email report, assessing its risk, determining whether it is malicious, and taking proportionate action. The goal is both to protect the person who reported it and to find related threats that may still be active elsewhere.
In Microsoft 365, employees can use Outlook's built-in Report phishing action. Depending on how administrators configure reporting, the message can go to Microsoft, an organization's reporting mailbox, or both. Security administrators can review messages on the User reported tab of the Defender Submissions page. See Microsoft's reporting settings and submission workflow.
Connect reports to tickets and reporter feedback. Do not ask employees to forward suspicious mail to colleagues for informal review: forwarding may lose useful message metadata and expose additional recipients. Use your approved reporting process instead.
The following seven-step sequence works whether your team investigates every submission manually or uses approved automation for the routine cases.
Step 1: Configure the Reporting Pipeline Before an Incident
A SOC cannot triage reports reliably if employees do not know how to submit them or reports disappear into an unmonitored mailbox.
In the Microsoft Defender portal, open Settings → Email & collaboration → User reported settings. Review whether Outlook's built-in Report button is enabled and where reports are sent. If your company uses a third-party reporting add-in, confirm that its messages reach the intended investigation queue and retain the evidence analysts need.
Then establish four operational rules:
- One intake queue: Route user submissions into a monitored workflow with ownership, ticket IDs, and escalation coverage.
- Clear priority labels: Use critical, high, normal, and low, with response targets based on risk rather than simply arrival order.
- Minimum evidence: Capture the reporter, reported time, original recipient, message identifiers, and whether the user clicked, downloaded, replied, or submitted information.
- Feedback ownership: Assign someone to notify the reporter when the case is closed, including when the email is legitimate.
Access also matters. The Submissions experience is more broadly available, while Threat Explorer and certain advanced remediation and automated investigation capabilities depend on Defender for Office 365 Plan 2. Real-time detections is available in Plan 1, with narrower capabilities. Confirm your entitlements and roles before writing a playbook that relies on a specific button.
Finally, test the process with an approved simulation. Verify that a submitted report creates the expected entry and reaches a person or automation that can act on it. An unowned queue is a security gap.
Step 2: Decide Whether This Is a Triage Case or an Active Incident
The first analyst question should be: Has the reported message already caused a harmful action?
Ask the reporter whether they opened a link, entered a password, approved an MFA request, downloaded or executed a file, replied with sensitive information, scanned a QR code, or changed any payment instructions. Avoid making assumptions from a read receipt or a single click log; verify what actually happened.
A practical priority model looks like this:
Adjust targets to your operating hours and risks.
When an account may be compromised, do not wait to finish email classification. Follow the identity-response process: evaluate sessions and sign-in activity, revoke sessions where warranted, reset credentials through approved procedures, require phishing-resistant reauthentication, and investigate mailbox forwarding rules and malicious OAuth grants. If funds may have moved, alert the finance and fraud-response teams so they can contact the bank immediately.
Link the ticket to the broader incident.
Step 3: Examine the Message, Sender, and Security Verdict
Once immediate risk is addressed, inspect the original message in Microsoft Defender. Start with the user submission and open its details. Where your licenses and permissions allow, open the Email entity page to examine the delivery record, detection history, authentication results, URLs, attachments, and post-delivery events.
Record the Network Message ID and Internet Message ID, along with the sender, recipient, subject, and received time. These identifiers help distinguish a specific message from lookalike messages with similar subjects. Microsoft's Email entity page reference explains the available evidence.
Read the identity evidence carefully
Compare the visible From address with the return path, reply-to address when available, sending domain, and message headers. Look for lookalike domains, display-name impersonation, newly introduced senders, and a reply address that would redirect the conversation.
Review SPF, DKIM, DMARC, and composite authentication—but do not treat a pass as a safety verdict. An attacker can authenticate mail from a domain they own, and a compromised supplier account may send fully authenticated malicious messages. Authentication supports a conclusion about origin; it does not establish that the request is trustworthy.
Reconcile the product verdict with the business context
Was the message delivered to the inbox, junked, blocked, quarantined, or removed after delivery? Did an allow rule, policy override, or user action change that outcome? Does Defender identify phishing, impersonation, malware, or no threat?
A verdict is evidence, not the investigation's entire conclusion. A message that asks an employee to change payroll or supplier bank details may be dangerous even when it contains no suspicious URL or attachment. Conversely, an unfamiliar sender and urgent wording do not automatically make a legitimate billing reminder malicious.
Record the requested action and whether it fits the relationship. Intent-based reasoning examines what the sender wants, not just technical indicators.
Step 4: Investigate Links, Attachments, and Interaction Evidence
Inspect suspicious content without exposing an analyst workstation or spreading the message. Use authorized security tools, sanitized previews, and isolated analysis environments. Do not open questionable attachments or enter credentials on a suspected website to “see what happens.”
For a URL-based lure, check the real destination, redirect chain where available, domain reputation, hosting context, and whether the page impersonates a familiar login or file-sharing service. Examine Safe Links and related click evidence when those protections are deployed. A click alone does not prove credential theft.
For attachments, record filenames, types, hashes where available, and detection details. Escalate suspicious macros, script files, password-protected archives, or unusual document-sharing flows according to policy. A PDF can hide a credential-harvesting link; unfamiliar formatting alone does not prove malice.
Inspect QR-code lures as carefully as ordinary links. A QR code printed inside a PDF or email image can hide the destination from a quick glance; use tools that decode it safely rather than scanning it with a personal phone.
For payload-free BEC, investigate the transaction request itself. See StrongestLayer's VEC detection guide for related scenarios. Has this vendor previously requested an account change? Does the invoice match the purchase order? Is the request coming from a familiar thread but introducing a new beneficiary? The safest response to payment changes is independent verification using a trusted phone number or vendor portal—not replying to the potentially compromised conversation.
If the employee interacted with the message, corroborate their account with available identity, endpoint, and web-protection telemetry. Preserve timestamps and distinguish what is known from what is inferred. This helps the incident responder avoid both overreacting to harmless clicks and missing credential or session compromise.
Step 5: Hunt for Related Messages Across the Organization
One report may represent dozens of unreported deliveries. Before closing the case, search for other recipients and related messages.
In Defender for Office 365 Plan 2, Threat Explorer lets analysts search and investigate email activity. Plan 1 organizations can use Real-time detections for more limited investigation. According to Microsoft's Threat Explorer guidance, the available views and actions differ by plan and permissions.
Begin narrowly with the Network Message ID, then expand using a combination of sender address, domain, subject fragments, URL, attachment indicators, delivery window, and campaign or clustering information. A subject-only match can catch legitimate messages; a domain-only match can pull in unrelated mail. Validate the resulting set before taking action.
A useful campaign-scope record captures:
- How many distinct recipients received a matching message?
- Where are those messages now: inbox, junk, quarantine, or deleted?
- Were some messages already removed by Zero-hour auto purge (ZAP) or another control?
- Did any recipients click, reply, authenticate, or perform the requested business action?
- Does the same infrastructure appear in messages with different subjects or sender names?
Where advanced hunting is available and authorized, use email telemetry such as EmailEvents, EmailUrlInfo, EmailAttachmentInfo, and EmailPostDeliveryEvents to pivot across identifiers. Match related evidence carefully by NetworkMessageId and recipient as appropriate. Do not assume that every email with a similar subject belongs to the same attack.
Microsoft describes both email hunting schema tables and post-delivery events. Without these features, use available message-trace, submission, and incident evidence.
The output is a validated scope. Containment must target defensibly related messages and accounts.
Step 6: Contain and Remediate, Then Verify the Outcome
Remediation should follow the case's risk and approved response authority. For a confirmed malicious campaign, typical actions include removing or quarantining delivered copies, blocking verified malicious indicators, reviewing access by exposed accounts, and escalating endpoint or identity concerns.
In Threat Explorer, authorized Plan 2 responders can use Take action to move or delete messages, submit evidence, or initiate other supported actions. Some operations require particular roles or approval. Microsoft documents these boundaries in its Threat Explorer remediation guide. Avoid hard-deleting evidence when your investigation, legal hold, retention, or incident policy requires preservation.
Microsoft's ZAP can also act on malicious messages after delivery. That is useful, but it does not remove the need to check actual mailbox state: outcomes differ by policy and the original message may have been read or interacted with before a post-delivery action occurred.
Before the ticket is marked contained, verify four facts:
- Message state: Affected copies are no longer accessible where removal was intended, or the remaining exposure is explicitly documented.
- Identity state: Any suspected account compromise has been investigated and necessary actions have been confirmed.
- Business state: Finance, HR, or other process owners have checked whether a fraudulent instruction was acted on.
- Control state: Any new blocks or policy changes are scoped appropriately and have not created unacceptable false positives.
Avoid broad blocks on shared providers or permanent allow entries without investigating the underlying verdict.
When evidence is inconclusive, record the case as unresolved or under monitoring. A defensible uncertain verdict is better than a confident but unsupported “safe.”
Step 7: Classify, Document, and Notify the Reporting User
Use consistent final dispositions:
Document the reporter, ticket owner, relevant message IDs, indicators, business intent, evidence reviewed, scope, user interactions, disposition, actions taken, verification, and closure time. Add a short narrative that another analyst can understand without opening every console.
Then respond to the reporter. Microsoft supports Mark as and notify for user-reported messages in the Submissions workflow. The message can tell the employee whether phishing was found or whether no threat was identified, depending on the admin review verdict.
Keep feedback specific: “Thanks for reporting this. We confirmed it was a phishing attempt and removed related messages from affected mailboxes. If you interacted with the link, contact the security team immediately.” For a legitimate message, explain that it was reviewed and encourage future reporting rather than making the employee feel embarrassed.
Close the ticket after containment and feedback; link continuing incidents.
A Realistic Triage Example: A Supplier Changes Banking Details
Imagine an accounts-payable employee reports an email in an existing supplier conversation. It says that the next invoice must be paid to a different bank account before the end of the day. The email has no link and no attachment. SPF, DKIM, and DMARC appear to pass.
Authentication and absence of payload do not settle the business risk.
Intake: The analyst confirms the employee has not made the payment and has not called the phone number in the email. The case is prioritized because it involves an imminent financial action.
Message analysis: The sender appears to be the real supplier. Authentication checks pass, but the request introduces new remittance instructions and unusual urgency. Neither observation proves compromise, so the analyst records the concern without claiming certainty.
Independent validation: Finance calls a verified vendor contact using details already on file. The vendor confirms that no bank change was requested. The team pauses the payment workflow and opens a fraud-related incident.
Campaign scoping: The SOC searches for similar requests and identifies another recipient in procurement. The second employee has not acted. Both messages are preserved and handled through approved remediation procedures.
Closure: The reporting employee receives confirmation, finance verifies the vendor master record, and the security team records the attempted payment diversion. The case is classified as a malicious business-email-compromise attempt based on the independent validation and message evidence.
The lesson: assess what the email asks the business to do—not just its sender and payload.
A One-Page Phishing Triage Checklist for SOC Analysts
Use the following sequence as a quick-reference aid during a live investigation.
Intake and priority
- Identify the reporter, affected mailbox, message, and time.
- Ask about clicks, credentials, attachments, replies, MFA actions, and payments.
- Escalate active compromise or financial exposure immediately.
Evidence and analysis
- Preserve message identifiers, sender information, headers, and current delivery state.
- Check Defender verdicts, detection history, policies, and overrides.
- Review URLs, attachments, QR codes, and relevant click evidence safely.
- Describe the requested action and test it against trusted business context.
Scope and response
- Search for related recipients and campaign variants.
- Confirm whether ZAP or prior administrative action already changed message state.
- Apply approved containment, blocking, and identity-response actions where warranted.
- Verify outcomes and preserve necessary evidence.
Closure
- Record a supported disposition and short case narrative.
- Notify the reporter; link any continuing incident.
- Update relevant detection rules, reporting guidance, or coaching based on the case.
How to Measure Phishing Triage Without Rewarding the Wrong Behavior
Measure security outcomes as well as ticket speed.
Start with five useful measures:
- Median time to first analyst assessment: Time from report receipt to the first documented risk evaluation. Use the median so a handful of long cases do not distort the typical experience.
- Median time to verified containment: Time from the initial report to confirmed removal or risk reduction for cases that required containment.
- Malicious-report yield: The share of reviewed user reports that are ultimately classified as malicious. Interpret alongside report volume.
- Repeat-recipient exposure: The number of additional affected mailboxes found after the first report, segmented by attack type.
- Reporter feedback completion: The percentage of closed reports for which the employee received a clear outcome.
Segment the figures by case severity, phishing type, business function, and whether automation completed the first assessment. Measure errors, reversals, and user-reported false positives as well. Faster triage must not increase misses or false positives.
Where Microsoft Defender for Office 365 Plan 2 and the relevant Security Copilot capabilities are licensed, Microsoft's automated investigation and phishing-triage features can help process some submissions. Use automation to reduce repetitive work while keeping human review for uncertain intent, significant business consequences, and exceptions.
Where StrongestLayer Fits in a Microsoft 365 Triage Workflow
Microsoft Defender supplies core reporting, investigation, and remediation tools. Complementary security should add useful context rather than another alert queue.
On its platform overview, StrongestLayer describes its TRACE reasoning engine around email intent and context, with AI Inbox Advisor providing explanations to users and Threat Triage supporting analyst review. Its published platform description also includes hunting for related email indicators and integration with Microsoft 365. Those capabilities are relevant when a suspicious request carries little technical payload but substantial business risk.
Evaluate with a compromised vendor account, QR-code lure, executive impersonation, and unusual legitimate email. Ask whether the results explain the verdict, preserve evidence, help identify related messages, and fit into existing SOC workflows. Verify specific integrations, response actions, and performance expectations during a proof of concept rather than assuming every environment behaves the same way.
The aim is faster, explainable decisions—not replacing analyst judgment.
Final Thoughts
The work begins when an employee presses Report phishing: establish risk, scope the campaign, contain exposure, and give useful feedback.
Microsoft 365 provides the reporting and investigation foundation. A clear runbook, trusted business verification, and well-governed automation turn that foundation into consistent decisions. Start with a repeatable seven-step process, test it with realistic scenarios, and measure what happens after the first report—not just how quickly the ticket disappears.
FAQ
Where do user-reported phishing emails appear in Microsoft 365?
Depending on your configuration, they are sent to Microsoft, a reporting mailbox, or both. Administrators can review them under the User reported tab of Submissions in the Microsoft Defender portal. The built-in Outlook Report button is configured under User reported settings.
Does Microsoft Defender automatically investigate reported phishing emails?
Microsoft Defender for Office 365 Plan 2 supports automated investigation and response playbooks for eligible alerts, including user-reported phishing. Actual behavior depends on alert policies, configuration, permissions, and licensing. Review the evidence and recommended actions before assuming the message or related threats have been contained.
Can phishing pass SPF, DKIM, and DMARC?
Yes. Those checks help establish whether sending infrastructure is authorized for a domain. A lookalike domain can be correctly authenticated, and a compromised legitimate mailbox can send authenticated malicious requests. Investigate message intent, sender relationships, and the requested action as well.
Should the SOC delete every email reported by employees?
No. Reports may involve spam, legitimate mail, simulations, or actual attacks. Verify the disposition and scope before taking destructive action. Follow preservation, retention, and approval requirements.
What if a user clicked the link before reporting?
Escalate according to the destination and observed activity. Check for credential entry, suspicious sign-ins, session misuse, malicious downloads, or business actions. A click alone does not prove compromise, but it may justify deeper investigation.
What is the difference between phishing triage and incident response?
Triage determines the nature, priority, and scope of a suspicious report. Incident response coordinates containment, investigation, recovery, and communications when malicious activity or business impact is confirmed or reasonably suspected. A triage case often becomes the first record of a larger incident.