Skip to main content

Nerdio is Microsoft’s 2026 Americas ISV Partner of the Year Get the story

Blog

How to detect and remediate policy drift across Microsoft 365 tenants

Detect and remediate policy drift across Microsoft 365 tenants with baselines, exception registers, monitoring, and audit trails.

A client's cyber insurer wants proof that multifactor authentication (MFA) is enforced for every admin. Yet a Conditional Access policy deployed eighteen months ago sits in report-only mode with three unexplained exclusions. That difference between the approved baseline and today's configuration is policy drift.

This guide is for managed service provider (MSP) owners running Microsoft 365 across tenants for dozens of clients. It also serves enterprise end-user computing (EUC) and governance, risk and compliance (GRC) leads managing several tenants after an acquisition or regional split.

What is policy drift?

Policy drift means the delta between the configuration you approved and the one enforced across Entra ID, Microsoft Intune, Exchange Online and the rest of Microsoft 365.

This guide uses policy drift for the security-control subset of what NIST and Microsoft call configuration drift. Neither source defines policy drift as a separate standards term, so the two overlap: every policy drift is a configuration drift, but not every configuration drift is policy drift.

NIST defines a baseline configuration as "A documented set of specifications for a system, or a configuration item within a system, that has been formally reviewed and agreed on at a given point in time, and which can be changed only through change control procedures". Microsoft adds: "When a delta exists between what a configuration baseline defines and the actual settings on a tenant, Tenant Governance reports a configuration drift."

The reason to call out policy drift separately is ownership. Configuration drift as a whole covers any technical deviation in infrastructure and tends to sit with DevOps or platform teams. Policy drift narrows that to deviations in the controls governing access and behavior, which security and GRC teams own.

In a Microsoft 365 tenant, a Conditional Access setting is both technical configuration and a security control, which is why it shows up on both teams' radar. The operational problem is a tenant that passes review in January and fails it in June without a change ticket, because device configurations have gradually moved out of their intended state.

Why tenants drift when nobody changes anything on purpose

Tenants drift because settings and exceptions change faster than anyone reviews them, and most of that change comes from documented Microsoft behavior rather than a rogue admin. Five mechanisms account for most of the drift across a multi-tenant estate:

  • Intune security baseline versions require deliberate updates:
    A new version replaces the previous one. Existing profiles stay on the old version with their settings read-only, and new settings don't apply until an admin edits and saves the profile. Across 48 tenants, each profile moves only when that tenant's admin acts.
  • Policy types overlap and report how conflicts resolve:
    Conflicts between Intune policy types can stop settings from applying, and duplicate compliance settings resolve to the most restrictive value. The Conflict status appears in the device configuration report, one tenant at a time.
  • Conditional Access exclusions can outlive their purpose:
    Exclude overrides include, and Microsoft recommends excluding at least one emergency access account from every policy. Microsoft manages exclusion lists per policy and per tenant, so an exclusion added for one client stays in place until someone removes it.
  • Microsoft stages managed policy rollouts:
    Microsoft-managed Conditional Access policies arrive in Report-only mode, and Microsoft enables them at least 30 days later unless admins act or opt out. The same rollout ends in a different state in each tenant, depending on what each admin did.
  • Access expiry and baseline reassignment require follow-up:
    A GDAP relationship lasts at most two years, and an expired one can't be restored. In Microsoft 365 Lighthouse, re-assigning a baseline overwrites tenant-specific exclusions, which then have to be re-entered per tenant.

Admin behavior adds to this. A help-desk exception nobody reverted and a vendor script that duplicated a policy push the baseline document further from the live tenant. The stale baseline goes unnoticed until someone outside IT asks you to prove it still holds.

What drift costs when the insurer or auditor asks

Drift costs you evidence, and the request for evidence arrives on someone else's schedule. In the 2026 State of the MSP Business Report, 49% of MSPs said insurers had asked them to prove standardization, while only 20% enforce one security standard across all tenants. Munich Re describes underwriting as moving "From self-assessment to independent evidence".

The Verizon 2026 Data Breach Investigations Report found that half of weak-password and permission-misconfiguration findings took almost eight months to resolve, against half of MFA findings resolved within one month (n=7,513). CIS Safeguard 4.1 requires a documented secure configuration process reviewed annually, and CIS says configuration "must be continually managed to avoid degrading security as software is updated or patched". The remediation cycle can therefore outlast the evidence window needed to explain it.

The free tier retains Entra audit logs for 7 days, and P1 and P2 retain them for 30 days, while CIS Safeguard 8.10 expects audit logs kept for at least 90 days. A screenshot of today's Conditional Access blade records the current tenant state without explaining how it got there.

How management platforms extend Microsoft's native drift detection

Microsoft's native tooling detects configuration drift inside a tenant and directs remediation through the administration experience that owns the resource. Management platforms extend those workflows across tenants.

Entra Tenant Governance and the TCM APIs

Tenant Governance, built on the TCM APIs, snapshots a tenant's configuration as a JSON baseline across over 200 resource types. Those resource types span Entra and Intune, plus Exchange Online and three other workloads.

Monitors check it every six hours and are set up in each tenant, capped at 30 monitors per tenant, so 48 tenants means 48 separate monitor setups. Microsoft's Entra blog announced Tenant Governance as generally available, though some Graph API reference pages are still marked beta. Microsoft also states that "Tenant Governance reports drift, but remediation happens in the administration experience that owns the drifted resource."

Detection follows a consistent cadence. Setup and remediation remain tenant-scoped.

Microsoft 365 Lighthouse

Lighthouse is free for eligible CSP partners and flags tasks that regressed from Compliant or Dismissed to Not compliant or Not licensed, with who, where and when details. The regression details tell the partner which tenant and task need follow-up.

Changes made outside Lighthouse are detected only at the next scheduled or manual scan, and deployment happens tenant by tenant. Multi-tenant reporting on deployments is "Planned for a future release".

Intune reports and Microsoft365DSC

Intune's security baseline monitoring reports per-setting Error and Conflict status, which Microsoft describes as deployment state rather than security state, with no multi-tenant roll-up. Devices check in every 8 hours by default, so a corrected policy reaches the device later than the console suggests. Microsoft365DSC checks every 15 minutes by default and ApplyAndAutocorrect reverts drift, but monitoring runs per host where the configuration was applied.

These tools cover drift within their tenant or host context. Nerdio Manager for MSP extends and enhances these Microsoft workflows by comparing assigned customer policies with an MSP-level source policy and letting admins fix, roll back or accept drift.

Cross-tenant detection and remediation is the mechanical half of the problem. The other half is knowing whether a drifted policy is an approved exception or an accident.

A remediation workflow that separates approved drift from accidents

Some drift is a documented exception, so reverting every delta breaks clients. The following seven steps give you a workflow for separating approved exceptions from accidental changes before you decide what to revert.

  1. Document the baseline and tier it:
    NIST CM-2 directs organizations to develop and document a current baseline configuration, then maintain it under configuration control. A practical tiering approach is to standardize common settings and document client-specific exceptions.
  2. Put changes behind approval:
    Intune Multi Admin Approval holds a change until another account approves it. That keeps unapproved settings out of enforcement. Report-only mode previews Conditional Access impact.
  3. Monitor continuously, review on a schedule:
    Six-hour monitors catch the delta. Perform monthly scans and quarterly deep dives. Use those reviews to catch expired exceptions.
  4. Validate before you revert:
    Establish who made the change and whether a ticket covers it. Intentional drift gets accepted with an expiry date. Unintentional drift gets the baseline reapplied.
  5. Register exceptions with an owner and an end date:
    Emergency access accounts stay excluded from policies that block sign-in; for other exclusions, exclusion groups can be governed through access reviews.
  6. Keep versions you can roll back to:
    NIST CM-2(3) calls for retaining an organization-defined number of previous baseline versions to support rollback. Deleted Conditional Access policies are recoverable for 30 days. Intune doesn't natively restore a deleted policy. Nerdio Manager bridges that gap by backing up and restoring Intune policies.
  7. Retain the trail past the audit window:
    Intune audit logs are kept for two years. Entra keeps them 7 or 30 days unless exported to Log Analytics.

Run end-to-end, these steps turn drift from an audit liability into a dated change record you can hand to an insurer or auditor.

How Nerdio Manager applies one baseline across every tenant

Nerdio Manager applies source policies across customer tenants and records the remediation decision for each drifted policy.

Solution baselines and Intune policy drift checks

Nerdio Manager allows you to define solution baselines. Teams use those baselines for initial deployments and to maintain consistent policy states across all customer accounts. Workloads include Entra ID, Intune, Exchange Online and Defender. Each Solution Baseline shows its status as Green (Valid/Satisfied), Yellow (Mismatch/Drifted), Red (Not Found/Drifted) or Gray (Excluded). Those statuses separate satisfied and excluded baselines from mismatches and missing policies.

A separate MSP-level drift check compares each customer's Intune policies against your source policy. It covers compliance and configuration policies, Conditional Access and update rings. Each drifted policy gets three actions:

  • Fix drift publishes the source policy back to the customer.
  • Rollback returns the policy to a specific earlier version.
  • Accept drift records the exception for 30, 60 or 90 days or indefinitely, with notes; the Accepted Drifts page shows the expiration date.

Nerdio Manager backs up policies daily and retains every previous version and its change log, giving admins a known version to restore after an accidental or unapproved change. A dashboard can show that a policy has drifted from the approved baseline, but it can't tell you whether the change was intentional, who approved it, or whether the exception still has a reason to exist.

What customers report

L7 Solutions reported a 90% reduction in IT setup time. Chris Brannon, their Director of Technology, said "Before Nerdio, setting up a new client took nearly a full workday."

Yanick Rose, Cloud Team Lead at MSIT, said that before, "we'd have a technician off the desk for two or three days just configuring baseline policies manually." They reported onboarding all 47 customers in two weeks and standardizing Intune and security policies across every tenant.

Where to start with the tenants you already manage

When the cyber insurer asks whether MFA is enforced for every admin, answer with a current baseline and exception register instead of scrambling through a Conditional Access policy left in report-only mode for eighteen months. Start by snapshotting the tenant most likely to get that insurer letter, using Tenant Governance or a platform baseline.

Then bring its Conditional Access exclusions into line with a register that has a named owner and end date per entry, and update any Intune baseline profile still on a superseded version.

In the opening scenario, the team would accept two exclusions indefinitely with notes after identifying them as break-glass accounts. The third belonged to a contractor who left, so the team would reapply the baseline. The insurer would receive a dated change log instead of a screenshot.

To see how solution baselines, MSP-level drift checks, and accepted-drift records work across every tenant you manage, get a demo of Nerdio Manager for MSP or walk through a self-guided demo.

Frequently asked questions about policy drift

Ready to get started?