Protego field desk
Identity & Access · 30% of exam

L2. Conditional Access Policies

Course outlineLesson 2 of 18

Conditional Access is the AZ-500 exam's most tested identity topic. Learn named locations, sign-in risk, MFA enforcement, and the policy evaluation order that determines whether access is granted or blocked.

Quick answer: Conditional Access is the Microsoft Entra ID policy engine that decides, at sign-in, whether to allow access, require extra controls such as MFA or a compliant device, or block. It needs Entra ID P1 (P2 for risk-based policies). Every matching policy applies, and any Block wins. Test new policies in report-only mode first. Exam update: AZ-500 retired on August 31, 2026. Conditional Access carries over to its replacement exam: see SC-500 lesson 1: PIM and Conditional Access.

What Is Conditional Access?

Conditional Access is Entra ID's policy engine that sits between identity verification and resource access. It evaluates signals including user identity, device state, location, and risk score, then enforces controls like requiring MFA, blocking access, or requiring a compliant device.

Think of it as an if/then policy: if these conditions match then enforce these controls.

Policy Anatomy

Every Conditional Access policy has three parts: Assignments cover who and what the policy applies to:

  • Users and groups (include/exclude)
  • Cloud apps or actions
  • Conditions: sign-in risk, user risk, device platform, locations, client apps
Access Controls define what happens when the policy matches:
  • Grant: block access, require MFA or an authentication strength, require compliant device, require hybrid Entra joined device
  • Session: sign-in frequency, persistent browser session, Conditional Access App Control (for example blocking downloads through Defender for Cloud Apps)

Licensing and Safe Testing

Conditional Access requires Microsoft Entra ID P1; risk-based policies require P2. Use report-only mode and the What If tool before enforcement, then review sign-in logs. Authentication strengths require a specified set of authentication methods, such as phishing-resistant MFA. The Authentication flows condition can target and block device code flow.

Security defaults blocks device code flow in new tenants starting July 1, 2026. Use Conditional Access when granular control and exceptions are required.

Named Locations

Named locations let you define trusted IP ranges or countries. Common exam pattern: create a named location for your corporate IP range, require MFA for sign-ins outside that location. A separate matching Block policy cannot be overridden by satisfying MFA. Exam trap: Named locations marked "trusted" reduce risk scores in Identity Protection. Marking an office IP range as trusted means sign-ins from that IP get a lower risk score.

Sign-in Risk vs User Risk

SignalSourceWhat It Measures
Sign-in riskReal-time and offline detectionsIs this specific sign-in suspicious?
User riskAggregated MLHas this account been compromised over time?
A policy requiring MFA for medium-plus sign-in risk catches anomalous sessions without blocking the user account. A policy requiring password reset for high user risk handles accounts with leaked credentials.

Build both as Conditional Access policies. The legacy user risk and sign-in risk policies inside ID Protection retired on October 1, 2026, and they were not converted automatically. For a step-by-step build of a baseline policy set, see Microsoft Entra Conditional Access setup.

Policy Evaluation Order

All matching policies are evaluated: there is no first-match-wins. If any matching policy blocks access, the user is blocked. Controls from grant policies are combined with AND logic by default, meaning the user must satisfy all controls. Exam tip: Exclusions are the only way to exempt a user from a policy. The Report-only mode lets you audit what a policy would do without enforcing it: use it before enabling new policies.

References

Exam Focus Points
  • ✓CA policy = Assignments (who/what/conditions) + Access Controls (grant/session)
  • ✓Named locations marked "trusted" lower the Identity Protection risk score for those IPs
  • ✓Sign-in risk = real-time per-session signal; User risk = aggregated account compromise signal
  • ✓All matching policies evaluate: no first-match-wins. Any blocking policy blocks the user
  • ✓Report-only mode audits policy impact without enforcement: use before enabling new policies
  • ✓Exclusions (not conditions) are the only way to exempt a specific user from a policy
Knowledge Check

1. A Conditional Access policy is set to 'Report-only' mode. What happens when a user triggers the policy conditions?

2. Which sign-in risk level is calculated in real time using Microsoft's threat intelligence and machine learning?

3. You want to exclude break-glass accounts from all Conditional Access policies. What is the correct approach?