A 30-second Conditional Access read and the four gaps it usually surfaces

A Conditional Access review does not need a week. Four controls carry most of the identity posture, and each one is a yes or no that Microsoft Graph answers directly. The ITSailor tenant scan reads the enabled policies and derives these four, then Copilot identity readiness on top. This is the same logic, described so you can run the read yourself.

Last verified: 2026-08-03.

Prerequisites

• Microsoft Graph access with Policy.Read.All.
• Global Reader or an equivalent read-only directory role is enough. No write access is needed.

The four controls, and what a gap means

1. MFA for admins. An enabled policy that targets directory roles or all users and grants on MFA. Missing this is a high gap: registering MFA is not the same as enforcing it at sign-in.
2. Legacy authentication blocked. An enabled policy that blocks Exchange ActiveSync and "other" clients. Legacy protocols bypass MFA and are the most common takeover vector, so a missing block is a high gap.
3. MFA for all users. A policy targeting all users with MFA. Missing this is a medium gap because coverage may be partial.
4. Device compliance gate. A policy requiring a compliant or hybrid-joined device. Missing this is a low gap and a candidate for sensitive apps.

If the tenant has no Conditional Access policies at all, that is a single high finding on its own: identity then rests on whatever Security Defaults provide.

Expected output

# Read enabled Conditional Access policies and their grant controls
Connect-MgGraph -Scopes "Policy.Read.All"
Get-MgIdentityConditionalAccessPolicy -All |
  Where-Object { $_.State -eq 'enabled' } |
  Select-Object DisplayName,
    @{n='Grants';e={ $_.GrantControls.BuiltInControls -join ',' }},
    @{n='ClientApps';e={ $_.Conditions.ClientAppTypes -join ',' }}

Read the output against the four controls above. The paid tenant scan does this automatically and adds a Copilot identity-readiness score at /tools/saas-auditor.

Side effects

This is a read. It changes nothing. The only cost is the Graph call and the reviewer time.

Rollback

No rollback applies to a read-only check. When you act on a gap, stage the new policy in report-only mode first, confirm it in sign-in logs, then enable it, so the enforcement itself has a rollback.

Limitations

This scores four controls, not every Conditional Access nuance such as named locations, risk-based policies, or session controls. Copilot identity readiness is only the identity side. The standing gate before any Copilot rollout is a data-oversharing and DLP review, which the score always flags and which this read does not perform.