Published: November 8, 2026 | By AzureWise Security | Microsoft Entra ID, Identity Protection, KQL, MFA Security

An attacker who steals a password may still be blocked by multifactor authentication (MFA). But what if the attacker gains access to an existing session and registers an authentication method they control? The next sign-in may look legitimate—even while the account is being taken over.

That is why authentication method changes deserve the same scrutiny as risky sign-ins and privileged role assignments. In October 2026, Microsoft documented a Suspicious credential addition detection in Microsoft Entra ID Protection, calculated in real time and designed to identify potentially malicious registration of new authentication methods. For security teams, this is an opportunity to tighten a frequently overlooked part of the identity attack chain.

This guide explains what the detection means, how to investigate suspicious registrations using Entra audit logs and KQL, and how to reduce the chance that a newly registered credential becomes an attacker’s foothold.

Why authentication method registration is an account takeover risk

A typical attack can unfold in four steps:

  1. An attacker acquires a password, phishes a session, or otherwise accesses a user account.
  2. The attacker reaches the authentication-method registration or management workflow.
  3. The attacker adds a phone number, authenticator method, or other credential that they control, if the tenant’s policies and available flows permit it.
  4. That credential may help the attacker regain or maintain access, even after the original entry point is discovered.

The important distinction: a successful MFA event does not prove that the registered method belongs to the legitimate user. Authentication method enrollment is itself a security-sensitive event.

Normal registrations happen every day: new employees enroll MFA, users replace phones, and help desks assist with recovery. The challenge is separating those ordinary changes from attacker-driven enrollment.

What changed in Microsoft Entra ID Protection?

Microsoft’s documentation now includes Suspicious credential addition, a real-time user-risk detection for a new authentication method added under suspicious circumstances. Microsoft lists Entra ID P2 as its license requirement. The detection can contribute to user risk and inform risk-based Conditional Access decisions, where those policies are configured.

The documented detection type is suspiciousCredentialAddition. This is an Identity Protection risk-event identifier, not an AuditLogs.ActivityDisplayName value. Do not paste it into an AuditLogs activity-name filter and expect matching results.

A detection is not the same thing as proof of compromise; conversely, a registration without a risk detection is not automatically safe. Security teams should combine Identity Protection signals with the audit trail and surrounding sign-in activity.

Step 1: Identify new authentication method activity in audit logs

If Microsoft Entra diagnostic settings route AuditLogs into a Log Analytics workspace, use this discovery query to find relevant activity in the last seven days.

AuditLogs
| where TimeGenerated > ago(7d)
| where ActivityDisplayName has_any (
    "security info", "authentication method", "authenticator"
)
| project TimeGenerated, ActivityDisplayName, Result,
          InitiatedBy, TargetResources, CorrelationId
| order by TimeGenerated desc

This is deliberately a discovery query, not an exact detector. Activity names vary by workflow, API, and operation. Inspect the actual event names and payloads in your tenant, then replace the broad matching logic with an explicit allowlist of the registration events you observe. Some matching events are removals, updates, or failed attempts—not additions.

If this query returns nothing, first confirm the workspace is receiving Entra audit logs and that the selected time range contains test activity. An empty result is not proof that no registrations occurred.

Step 2: Extract the actor and affected user

Once you have verified relevant event names, this query brings forward key investigation fields. Use the result to distinguish user-initiated changes from administrator-assisted or application-initiated activity.

AuditLogs
| where TimeGenerated > ago(7d)
| where ActivityDisplayName has_any (
    "security info", "authentication method", "authenticator"
)
| extend Actor = coalesce(
    tostring(InitiatedBy.user.userPrincipalName),
    tostring(InitiatedBy.app.displayName),
    Identity
)
| mv-expand Target = TargetResources
| extend TargetUser = tostring(Target.userPrincipalName),
         TargetName = tostring(Target.displayName)
| project TimeGenerated, ActivityDisplayName, Result,
          Actor, TargetUser, TargetName, CorrelationId
| order by TimeGenerated desc

Review the original TargetResources and InitiatedBy objects before relying on the extracted fields. Some events populate different target fields, and mv-expand can produce multiple rows per event. Treat the output as an investigation starting point—not as a canonical count of distinct registrations.

Prioritize events involving privileged users, unexpected administrators, changes outside established help-desk processes, or several registration changes in a short period.

Step 3: Correlate registration events with sign-ins

A method change becomes more concerning when it follows a sign-in from an unfamiliar IP address, an unexpected location, a new device, or a sign-in flagged for risk.

For a selected account, query SigninLogs around the registration timestamp:

let account = "user@contoso.com";
SigninLogs
| where TimeGenerated > ago(7d)
| where UserPrincipalName =~ account
| project TimeGenerated, UserPrincipalName, IPAddress,
          AppDisplayName, ResultType, RiskLevelDuringSignIn,
          ConditionalAccessStatus, DeviceDetail, LocationDetails
| order by TimeGenerated desc

Replace the placeholder UPN with the actual account. Compare sign-in timestamps to the audit event, inspect authentication details in the portal, and assess whether the sign-in was performed from an expected device and location. Sign-in log availability and risk fields depend on tenant configuration, licensing, and retention. A clean sign-in record does not eliminate the possibility of token theft or session abuse.

Step 4: Check Identity Protection and decide on containment

In the Microsoft Entra admin center, review ID Protection → Risk detections and Risky users for the affected account. Look for Suspicious credential addition and other correlated signals. Examine available detection details and the relevant sign-ins before deciding whether the action was authorized.

For potentially compromised accounts, follow your organization’s incident-response runbook. Depending on evidence and authority, actions can include:

  • Verify the change using a known-good channel—not a contact method that may have just been added.
  • Revoke active sign-in sessions and reset credentials where appropriate.
  • Remove unauthorized authentication methods after preserving the relevant audit evidence.
  • Check for additional persistence, including app consent grants, mailbox forwarding rules, and privileged role changes.
  • Restore the user’s approved methods through a controlled identity-verification and recovery process.

Be careful not to remove a user’s only legitimate recovery path without coordinating a secure replacement. For administrators and emergency-access accounts, follow the organization’s specialized procedures.

Step 5: Harden registration before the next incident

Detection is only half the solution. Reducing the ways an attacker can register a new credential is equally important.

Protect registration with Conditional Access. Review policies for the Register security information user action, use appropriate trusted conditions and strong authentication controls, and test changes in report-only mode where supported. Validate user onboarding, device enrollment, and recovery scenarios before broad enforcement.

Reduce weak and unnecessary methods. Review the Authentication methods policy. Prefer phishing-resistant authentication for users and roles where supported, and restrict methods that are no longer needed. Plan exceptions and accessibility needs rather than imposing a change without a recovery path.

Establish a monitored recovery process. Require identity verification for help-desk-assisted resets and method replacements. Record who authorized a change, what was changed, and why.

Alert on high-value events. Build scheduled alerting for authentication-method changes to privileged accounts, changes by unexpected actors, or bursts of related events. A Log Analytics alert can supplement—but should not be described as replacing—Microsoft’s native Identity Protection detection.

Review your coverage. Confirm Entra audit logs reach the intended workspace, set retention appropriate to your response requirements, and periodically test that registration activity appears as expected.

A practical checklist for security teams

  • [ ] Confirm whether Entra ID P2 and the Suspicious credential addition detection are available in your tenant.
  • [ ] Verify that Entra AuditLogs are collected and searchable.
  • [ ] Identify the exact authentication-method event names used by your tenant.
  • [ ] Monitor privileged-user registration changes and suspicious registration patterns.
  • [ ] Correlate changes with sign-ins, user risk, device details, and help-desk records.
  • [ ] Review Conditional Access protection for security-information registration.
  • [ ] Test investigation and secure recovery with a controlled test account.

Final thoughts

Identity security does not end when a user completes MFA. The moment an authentication method is enrolled, replaced, or removed is also a trust decision—and a potential account takeover opportunity.

Microsoft’s Suspicious credential addition signal gives Entra ID Protection customers another way to detect abuse. Pair it with properly collected audit logs, thoughtful KQL investigations, and carefully designed registration controls to make this identity attack path harder to exploit.

Explore AzureWise: The AzureWise Identity Security Assessment helps teams review Entra identity security signals, including risky sign-ins, MFA posture, and privileged changes. The KQL examples in this article are educational starting points; verify and tailor them before using them in production detection rules.

References

Put the guidance into practice

Want the Azure identity assessment already built?

AzureWise Identity Security Assessment turns Microsoft Entra ID telemetry into a weighted Security Rating, detailed findings, and prioritized remediation guidance inside Azure Monitor.

Get AzureWise — $99