Skip to content
Pro IT NW

Blog · 9 min read ·

Share

Entra ID Protection risk policies retire October 1, 2026

Legacy Entra ID Protection risk policies retire October 1, 2026. The replacement is a Conditional Access policy — and Microsoft's license table requires Entra ID P2 for both, which Microsoft 365 E3 and Business Premium do not include.

There is a small, quiet deadline sitting in the last week of September and the first day of October that most mid-market tenants have not put on a change calendar. On October 1, 2026, Microsoft retires the legacy risk policies configured inside Entra ID Protection. The guidance is unambiguous, and it is a Warning box rather than a footnote: "The legacy risk policies configured in Microsoft Entra ID Protection will be retired on October 1, 2026."

This post does three things: tells you in about a minute whether it applies to you at all, sets out what the replacement actually is, and names the two traps that turn a twenty-minute change into a rescheduled one.

The one-sentence version: if your tenant has Entra ID P2 and somebody once turned on the user risk or sign-in risk policy inside ID Protection, you have a Conditional Access policy to build before October 1 — and the admin role that builds it is not allowed to switch off the old one.

First, the question that decides whether you can stop reading

Do you have Entra ID P2? Everything below depends on it, and a large share of the regulated mid-market does not.

Microsoft's license table for ID Protection is explicit. Against the capability "Risk policies — Sign-in and user risk policies (via ID Protection or Conditional Access)", the three columns read Microsoft Entra ID Free: No, Microsoft Entra ID P1: No, Microsoft Entra ID P2 / Microsoft Entra Suite: Yes. The page opens the section with "Using this feature requires Microsoft Entra ID P2 licenses."

Now map that onto what a mid-market tenant usually owns. Microsoft's own pricing page states that Entra ID P1 "is available as a standalone or included with Microsoft 365 E3 for enterprise customers and Microsoft 365 Business Premium for small to medium businesses", and that P2 "is available as a standalone or included with Microsoft 365 E5 for enterprise customers."

The 60-second triage. If your users are licensed with Microsoft 365 Business Premium or E3 and you have not bought an Entra ID P2 or Entra Suite add-on, then you never had these policies, there is nothing to migrate, and this deadline is not yours. If you are on E5, or you bought P2 for a subset of users, keep reading. Check it in the Entra admin center under licenses rather than from memory — mixed licensing across a tenant is the norm in this size band, not the exception.

Worth being precise about one thing, because the license line is easy to over-read: a P1 tenant is not blind to identity risk. Microsoft's table gives Free and P1 "Limited Information. Only users with medium and high risk are shown. No details drawer or risk history" on the risky users report. You can see that something happened. What P1 cannot do is act on it automatically. That distinction is the whole product boundary, and it is the honest answer to "should we buy P2" — you are buying automated response, not visibility.

What is actually retiring, and what is not

ID Protection is not going away. The detections keep running, the risky users and risky sign-ins reports stay, the Graph APIs stay, and the Sentinel and Defender XDR integrations stay. What retires is the older configuration surface — the user risk policy and sign-in risk policy you set inside ID Protection itself, with an Enforce policy toggle.

The replacement is the same idea expressed as a Conditional Access policy, using User risk and Sign-in risk as conditions. Functionally you are moving the enforcement decision from a product-specific page into the tenant's general access-control engine, which is where the rest of your policy already lives. That is a good change. It is just one somebody has to make.

The migration, in the order Microsoft specifies

  1. Build the equivalent policies in Conditional Access, in report-only mode. Microsoft's recommended configuration is a user risk policy requiring risk remediation when user risk is High, and a sign-in risk policy requiring multifactor authentication when sign-in risk is Medium or High. Both start in report-only.
  2. Confirm the results, then move the toggle from Report-only to On.
  3. Disable the old policies in ID Protection — ID Protection, then Dashboard, then the User risk or Sign-in risk policy, then set Enforce policy to Disabled.

Note that the new policy is validated before the old one is turned off, so for a period both exist. That is deliberate and it is the right order. It also means the job is not finished when the Conditional Access policy goes live — step three is a separate action, and it is the one most likely to be forgotten once the new policy is working.

Trap one: the role that builds the replacement cannot retire the original

Microsoft's role table for ID Protection is worth reading closely before you assign the work. Against Conditional Access Administrator, the "Can do" column reads "Create policies that factor in user or sign-in risk as a condition." The "Can't do" column reads "Read or write to legacy ID Protection policies."

So the administrator who can perform step one cannot perform step three. Full access to ID Protection sits with the Security Administrator role. In a mid-market tenant, where role assignment is often narrower than people assume and the person doing identity work holds exactly the role they needed last time, this is the detail that turns a scheduled change into a half-finished one: the new policy goes live, the old policy stays enforced because nobody in the room can disable it, and both are running until someone notices.

Decide before the change window which named person holds which role. If you are running least-privilege properly — and if you are reading this you probably are, because that is the same discipline behind a Tier 0 model — then this is a two-person change, and that is fine as long as both people are booked.

Trap two: the users who cannot self-remediate get blocked

This is the one that generates helpdesk tickets, and Microsoft flags it as a Warning: "Users must register for Microsoft Entra multifactor authentication before they face a situation requiring remediation. For hybrid users that are synced from on-premises, password writeback must be enabled. Users not registered are blocked and require administrator intervention."

Read the last sentence again. The failure mode of a risk policy meeting an unregistered user is not a prompt — it is a block plus a call to your service desk. In a hybrid tenant, password writeback is the specific dependency, and it is exactly the kind of setting that was configured once during the original Entra Connect deployment and has not been looked at since. If you are already touching hybrid identity this month because Entra Connect Sync stops working for out-of-date versions on September 30, 2026, check writeback while you are in there. The two deadlines are a day apart.

Two more configuration details from the same guidance, both easy to get wrong:

  • Do not put both risk conditions in one policy. Microsoft's Warning is explicit: "Don't combine sign-in risk and user risk conditions in the same Conditional Access policy. Create separate policies for each risk condition." Two policies, not one.
  • Exclude your break-glass and service accounts. Microsoft recommends excluding emergency access accounts "to prevent lockout due to policy misconfiguration", and service accounts and service principals — naming the Microsoft Entra Connect Sync Account specifically. Locking yourself out of the tenant with a risk policy is a real outcome, and the exclusion costs nothing to add.

One behavioral note that surprises people: choosing Require risk remediation automatically applies two further controls — Require authentication strength as a grant control, and Sign-in frequency — Every time as a mandatory session control. And for passwordless users, Microsoft revokes the session so they must reauthenticate, rather than prompting for a password change. If your organization has been moving to passkeys, the user experience of remediation is not the same as it was.

Why this is worth a change window rather than a note in a backlog

Risk-based Conditional Access is not an obscure feature. It is one of the specific controls that gets asserted, by name, on a cyber-insurance application, in a HIPAA security narrative, and in the identity section of a CMMC or NIST 800-171 control description. The gap between "we have automated response to risky sign-ins" and "we had automated response to risky sign-ins until an October retirement we did not action" is the kind of gap that is invisible day to day and awkward at renewal or assessment.

Microsoft's own behavior is a signal about volume here. The guidance publishes a dedicated support path for this one migration: raise a request described as "Migrate legacy ID Protection policy", under service type Microsoft Entra Sign-in and Multifactor Authentication, problem type Identity Protection, subtype Configure risk policies. Vendors do not build a bespoke intake route for a migration they expect three people to perform.

Our read, and it is ours rather than Microsoft's: Microsoft states that the policies are retired on October 1 and does not publish the post-retirement failure behavior. We are not going to assert that enforcement stops silently, because Microsoft has not said so. What we would plan around is the conservative version — do not assume an ID Protection policy is still enforcing after that date, and prove it either way from a report-only Conditional Access policy's own output rather than from an absence of complaints. An identity control that stops working does not necessarily announce itself.

What we would do this week

  1. Establish whether you hold Entra ID P2, per user, in the Entra admin center. If not, close the tab — this is not your deadline.
  2. Look in ID Protection for an enabled user risk or sign-in risk policy. Plenty of P2 tenants never turned these on; if both are disabled you have nothing to migrate either, and that is worth recording as a checked fact rather than an assumption.
  3. Confirm MFA registration coverage and, for hybrid tenants, password writeback before any policy goes to On. This is the step that decides how many blocked users you create.
  4. Build both Conditional Access policies in report-only — separately, with break-glass and sync-account exclusions — and read the report-only results for a few days.
  5. Book the two roles. Conditional Access Administrator to enable, Security Administrator to disable the legacy policy. Same change window.
  6. Write down what you checked and when. If a questionnaire or an assessor asks next year whether risk-based access control was continuously in place, the dated record is the answer.

Related reading

Source

  • Risk policies — Microsoft Entra ID Protection, Microsoft Learn (page date October 30, 2025; last updated April 28, 2026). Source for the retirement date, the migration steps, Microsoft's recommended risk levels, the MFA-registration and password-writeback warning, the exclusion guidance, and the dedicated support path.
  • What is Microsoft Entra ID Protection?, Microsoft Learn (page date October 30, 2025; last updated February 10, 2026). Source for the license table and the required-roles table, including the Conditional Access Administrator limitation.
  • Microsoft Entra plans and pricing, Microsoft. Source for which Microsoft 365 suites include Entra ID P1 versus P2.
The 30-second version. Legacy Entra ID Protection risk policies retire October 1, 2026. They require Entra ID P2, which comes with Microsoft 365 E5 but not with E3 or Business Premium — so check that first, because for many mid-market tenants this is not a deadline at all. If it is yours: build two separate Conditional Access policies in report-only, confirm MFA registration and hybrid password writeback, exclude break-glass and sync accounts, and book both a Conditional Access Administrator and a Security Administrator, because the first cannot disable what the second must.

If you would rather have this run as a bounded piece of work than added to an already-full queue, scope it with us — license check, report-only build, remediation-coverage review and a dated record of what changed.

Questions we get asked

What exactly retires on October 1, 2026?
The user risk and sign-in risk policies configured inside Microsoft Entra ID Protection. Microsoft's guidance carries a Warning box that reads, verbatim: 'The legacy risk policies configured in Microsoft Entra ID Protection will be retired on October 1, 2026.' ID Protection itself is not retiring — the detections, the risky users and risky sign-ins reports, and the Graph APIs all continue. What goes away is the older place you configured an automatic response to risk. That response moves to Conditional Access.
How do I know in 60 seconds whether this affects us?
Check whether you hold Entra ID P2. Microsoft's license table for ID Protection lists 'Risk policies — sign-in and user risk policies (via ID Protection or Conditional Access)' as No for Microsoft Entra ID Free, No for Entra ID P1, and Yes only for Entra ID P2 or Microsoft Entra Suite. Microsoft's pricing page states that P1 is 'included with Microsoft 365 E3 for enterprise customers and Microsoft 365 Business Premium for small to medium businesses', while P2 is 'included with Microsoft 365 E5'. So if your tenant runs on Business Premium or E3 with no P2 add-on, you never had these policies and there is nothing to migrate.
What replaces them?
Equivalent user risk and sign-in risk policies built in Conditional Access. Microsoft's migration path is three steps: create the equivalent policies in Conditional Access in report-only mode, confirm the results and move the toggle from Report-only to On, then disable the old policies in ID Protection. Note the ordering — the new policy goes in first and gets validated before the old one is switched off.
Which admin role can actually perform the migration?
This is the part that catches teams out, and it comes from Microsoft's own role table. A Conditional Access Administrator can 'create policies that factor in user or sign-in risk as a condition' but explicitly cannot 'read or write to legacy ID Protection policies.' So the role that builds the replacement cannot turn off the thing being replaced. You either need a Security Administrator, who has full access to ID Protection, or you need two people coordinating two steps. Work that out before the change window, not during it.
What happens if we do nothing?
Microsoft states that the policies are retired on that date and does not publish the failure behavior in detail, so we will not invent one. What we would plan around is this: after retirement, an automatic response to identity risk that you configured in ID Protection should not be assumed to still be running. That matters because risk-based access control is a control organizations routinely assert on cyber-insurance questionnaires and in compliance narratives. The safe move is not to guess — build the Conditional Access policy in report-only mode now and confirm from its output that risk is still being evaluated and acted on.

Written by the team at · Senior-led Microsoft project consultancy · Seattle and the Pacific Northwest, delivered USA-wide.

Have a project on the runway?

Tell us the workload, the seat count, and the deadline. We'll come back with scope and a fixed-fee range.