top of page

Universal Continuous Access Evaluation How Entra ID and GSA Close Access Control Gaps

Kostas Tsiolas
2 days ago
9 min read

A sign-in is only a snapshot. It proves that a user and device met access requirements at one moment, but it does not prove they are still trustworthy 10 minutes later.


That gap matters. A user account can be disabled. A device can fall out of compliance. A session token can be stolen. A user risk level can change after sign-in. Traditional access control often waits for a token to expire, for a new sign-in, or for an application to recheck policy. Attackers can use that delay.


Microsoft Entra ID already addresses part of this problem with Continuous Access Evaluation, or CAE, for supported applications. The challenge is that not every application understands CAE natively. That is where Global Secure Access, or GSA, changes the model.


Universal Continuous Access Evaluation connects Microsoft Entra ID with Global Secure Access so organisations can enforce access decisions more continuously, even for applications that do not have built-in CAE support. Instead of trusting only the sign-in event, access becomes a living decision based on current user, device, session, and policy signals.


Eye-level view of a rugged tablet showing a generic access status screen beside a hardware security key.
Access decisions need current signals, not only a past sign-in.

Why sign-in is no longer enough


For many years, access control centred on authentication. A user entered a password, completed multifactor authentication, met Conditional Access rules, and received a token. The application accepted that token until it expired or the user signed out.


That model is simple, but it has a weak point: trust can change mid-session.


Common examples include:


  • A user account is disabled after suspicious activity.

  • A device loses compliance because encryption is turned off.

  • A user’s risk level increases after identity protection signals detect unusual behaviour.

  • An administrator changes a Conditional Access policy.

  • An employee leaves the organisation while still holding an active session.

  • A token is copied from a compromised endpoint.


In a traditional access model, the application may not know about these changes until the next token refresh or sign-in. Some apps shorten token lifetimes to reduce the window, but that creates user friction and still does not provide immediate response.


Continuous evaluation shifts the question from “Was this request valid when the user signed in?” to “Is this request still valid now?”


That distinction is critical for zero trust. Zero trust does not mean blocking everything. It means granting access based on explicit verification and then continuing to verify as conditions change. Identity, device health, location, risk, and application sensitivity all become part of the decision.


Microsoft Entra ID plays the identity and policy role in this model. It receives signals, evaluates Conditional Access policies, and can respond to important events. Where applications support CAE natively, they can react to changes quickly. Where they do not, organisations need another enforcement point.


How Entra ID and Global Secure Access work together


Microsoft Entra ID is the control plane for identity-based access. It manages users, groups, authentication methods, risk signals, and Conditional Access policies. Global Secure Access adds a network access layer that can apply Entra-based controls to traffic moving towards applications and resources.


This matters because many applications were not built to understand CAE events. They may accept a valid token without checking whether the user, device, or session still meets current policy. Rewriting or replacing those apps is rarely practical, especially for older SaaS tools, private applications, or line-of-business systems.


Global Secure Access gives organisations another place to enforce decisions. Rather than relying only on the application to interpret CAE signals, GSA can sit in the access path and apply policy based on Entra ID’s current assessment.


In practice, the model looks like this:


Access model

Where the decision happens

Main limitation

Traditional sign-in

At authentication and token issuance

Trust may become stale during the session

Native CAE

In CAE-aware applications and services

Only works when the app supports CAE

Universal CAE with GSA

Through Entra ID policies enforced in the access path

Requires traffic to flow through Global Secure Access


With Universal CAE, Entra ID and GSA work together to reduce the gap between a policy change or risk event and the enforcement of that change. If a session is no longer valid under current policy, Global Secure Access can help deny or interrupt access even when the target application lacks native CAE behaviour.


This is especially useful for organisations that already have a mix of application types:


  • Microsoft 365 services with native CAE support.

  • Modern SaaS applications using Entra ID for sign-in.

  • Older web applications that support single sign-on but not real-time policy response.

  • Private applications accessed by remote or hybrid workers.

  • Internal tools that cannot easily be changed.


The benefit is consistency. Security teams can define policy in Entra ID and use GSA to extend enforcement beyond the subset of apps that support CAE directly.


What Universal CAE changes for access control


Universal CAE is significant because it closes a common blind spot: the period after successful authentication but before the next policy check.


That blind spot is not always obvious. Dashboards may show that Conditional Access is working because users are challenged correctly at sign-in. Yet an active session can remain usable after the facts change. The account may no longer be trusted, but the application may still accept the token.


Universal CAE reduces that problem in several ways.


It makes trust more dynamic


Access decisions become less dependent on token lifetime alone. If a user risk signal changes, a device becomes non-compliant, or an account is disabled, enforcement can respond closer to the time of the event.


This does not remove the need for good token design, session controls, or application security. It gives organisations a stronger way to respond when trust changes during active access.


It extends policy to more applications


Native CAE is powerful, but it depends on application support. Many organisations cannot wait for every application vendor or internal development team to support CAE. Universal CAE provides a path to apply stronger enforcement to applications that would otherwise lag behind.


That is important for environments with older systems. Legacy applications often hold sensitive data, yet they may have the weakest session controls. GSA can help protect access to those systems without demanding a full application rewrite.


It reduces reliance on broad network access


Traditional remote access often grants users wide access to network ranges once connected. Global Secure Access supports a more targeted model, where access can be tied to the user, device, application, and policy state.


That helps move away from “connected means trusted.” The access path can check whether the request still aligns with policy, rather than assuming that a previously authenticated connection is enough.


It supports faster response to security events


Security teams often need to act quickly during an incident. Disabling an account, raising user risk, changing policy, or marking a device non-compliant should have practical effect as soon as possible. Universal CAE helps translate those identity changes into access enforcement for more applications.


The real value of continuous access evaluation is not only better sign-in security. It is the ability to respond when trust changes after sign-in.

Wide-angle view of a secure data corridor with illuminated server racks behind locked mesh panels.
Identity policy becomes stronger when enforcement reaches the access path.

Real-world examples of Universal CAE in action


The strongest case for Universal CAE becomes clear when looking at everyday security scenarios. These are not rare edge cases. They are the kinds of access-control gaps that appear in many organisations.


A stolen token is used after a device is compromised


Imagine a user signs in from a managed laptop and passes MFA. The user accesses a private business application that does not support native CAE. Later, malware on the laptop steals a session token or changes the device state. Endpoint tools detect the problem and the device falls out of compliance.


Without continuous enforcement, the application may continue accepting the session until it expires. If the token is usable, an attacker may still have time to access data.


With Entra ID and Global Secure Access working together, the device compliance change can affect access decisions more quickly. GSA can help enforce the updated policy for the application, even if the application itself does not understand CAE.


The result is a shorter window of exposure.


A contractor’s account is disabled during offboarding


Contractors often need access to internal tools, document repositories, or partner systems. When a contract ends, the account should lose access quickly. In a traditional model, disabling the account may stop new sign-ins, but existing sessions can sometimes remain active until tokens expire or applications recheck identity state.


Universal CAE helps close that gap. Once Entra ID reflects the account change, GSA can enforce the loss of trust for protected applications in the access path. This is especially useful for private apps that are not frequently updated or do not have advanced session awareness.


For organisations with regular onboarding and offboarding cycles, this reduces the chance that access survives beyond the approved period.


User risk increases after an unusual sign-in pattern


A user signs in successfully from Greece in the morning. Later, identity protection signals detect activity that suggests impossible travel, unfamiliar behaviour, or another risk condition. The user risk level changes after the original sign-in.


For CAE-aware applications, Entra ID can support a faster response. For applications without native CAE, Universal CAE helps bring similar enforcement through GSA. Access can be blocked, interrupted, or re-evaluated based on Conditional Access rules and present risk.


This gives security teams a better way to handle identity risk during active sessions, not only at the next login prompt.


A policy change must take effect quickly


Security teams sometimes change Conditional Access policies in response to a new threat. For example, they may require compliant devices for a sensitive application, block access from certain conditions, or require stronger authentication.


With traditional enforcement, some active sessions may continue under the old assumptions. That creates a delay between “policy changed” and “access changed.”


Universal CAE helps reduce that delay for applications protected through Global Secure Access. The policy decision in Entra ID becomes more meaningful because the enforcement point can act on it during ongoing access.


A private application has sensitive data but limited security features


Many internal applications were built before modern identity controls became standard. They may support SSO, but not continuous evaluation, claim challenges, or fine-grained session control.


Replacing the app may take years. Leaving it exposed is risky.


GSA can help place that application behind identity-aware access controls. When paired with Entra ID Conditional Access and Universal CAE, the organisation gains stronger control over who can reach the app and under what current conditions.


This is often one of the most practical uses of Universal CAE: improving security around important systems that cannot be modernised quickly.


What organisations should plan before adopting it


Universal CAE is not a switch to flip without preparation. Its value depends on policy quality, traffic coverage, device posture, and operational readiness.


A good starting point is to identify where access-control gaps exist today. Look for applications that meet one or more of these conditions:


  • They contain sensitive data.

  • They lack native CAE support.

  • They use long-lived sessions.

  • They are accessed by remote users or third parties.

  • They rely on older authentication or network access patterns.

  • They are hard to modify or replace.


Next, review Conditional Access policies. Universal CAE will only be as useful as the signals and rules behind it. If policies are too broad, inconsistent, or based only on sign-in location, enforcement may not improve much. Stronger policies often include device compliance, user risk, sign-in risk, authentication strength, and application sensitivity.


Operational teams should also plan how to handle user experience. Continuous enforcement can interrupt access when conditions change. That is a security benefit, but users and support teams need clear messages and processes. A blocked session should lead to a useful path, such as reauthenticating, registering a compliant device, or contacting support.


Monitoring is equally important. Teams should review sign-in logs, Conditional Access results, GSA traffic visibility, and incident records. The goal is to confirm that policies work as intended and do not block legitimate work without reason.


A phased approach often works best:


  1. Start with a small set of sensitive applications.

  2. Apply policies in report-only or controlled test modes where available.

  3. Validate device and identity signals.

  4. Expand enforcement to more applications.

  5. Review exceptions regularly and remove them when they are no longer needed.


The key is to treat Universal CAE as part of a wider access strategy. It complements MFA, device management, identity governance, endpoint protection, and application security. It does not replace them.


Close-up of a hand holding a security key near a laptop with a blurred access prompt on screen.
Strong access control combines identity, device trust, and real-time enforcement.

The practical takeaway


Access control cannot stop at sign-in. The conditions that make a session trustworthy can change at any point, and security tools need to respond while access is still active.


Microsoft Entra ID provides the identity signals and Conditional Access policies. Global Secure Access provides an enforcement layer that can apply those decisions to more traffic and more applications. Together, they help address one of the most persistent access-control gaps: applications that do not support native CAE but still need better protection.


For organisations with hybrid work, cloud apps, private apps, contractors, and legacy systems, Universal CAE offers a practical path forward. It brings continuous trust assessment closer to the resources that need protection.


The most useful next step is to map high-value applications against their current session behaviour. If an app cannot react when a user, device, or risk state changes, it belongs on the shortlist for stronger protection through Entra ID and Global Secure Access.


Comments


bottom of page