Join our Newsletter — 33% off our NHI Course
Home› FAQ› Architecture & Implementation› How should organisations layer identity controls when Microsoft…
Architecture & Implementation

How should organisations layer identity controls when Microsoft Entra ID does not cover every legacy, OT, or on-premises system?

← Back to all FAQ
By NHI Mgmt Group Editorial Team Updated October 30, 2026 Domain: Architecture & Implementation

Organisations should treat Entra ID as one control layer, not the entire identity perimeter. A practical model combines modern MFA, passwordless authentication, and a separate security layer for legacy, OT, and off-network resources. The goal is to maintain access continuity while applying phishing-resistant assurance wherever Microsoft-native coverage is incomplete.

Layering Entra ID with Legacy and OT Identity Controls

Entra ID is strongest as a modern control plane, but legacy, OT, and air-gapped or on-premises systems often need separate authentication paths, local account governance, or brokered access. The practical design goal is to avoid a single identity assumption across the whole estate while still pushing phishing-resistant controls where they are supported.

That usually means defining which systems are authoritative in Entra ID, which remain local, and which are accessed through an intermediary such as a jump host, PAM workflow, or brokered remote access pattern. Where Entra ID coverage is incomplete, the risk is not just weaker sign-in strength, it is fragmented policy enforcement, duplicated accounts, and inconsistent revocation.

Why the Identity Perimeter Splits Across Modern and Legacy Systems

Legacy and OT environments often fail on the same point: the platform can keep running without a direct tie to cloud identity services, so organisations leave local logons, shared operator accounts, or vendor access in place longer than intended. That is why a layered model matters, and why an identity architecture for OT environments needs to treat brokered access, offline logon, and separate identity stores as design choices rather than exceptions.

In hybrid estates, the control question is not whether Entra ID exists, but where it stops being the system of record for access decisions. Once that boundary is clear, organisations can apply stronger assurance to the systems Entra ID does cover, while setting compensating controls for equipment that cannot consume modern federation, conditional access, or passwordless flows.

A second boundary is operational: some resources are always online and centrally managed, while others are uptime-sensitive, intermittently connected, or maintained by vendors. Those systems often need a different authentication cadence, different logging, and tighter segregation between administrative access and normal user access.

How to Build a Practical Multi-Layer Identity Model

The cleanest model is to separate the identity estate into three layers: Entra-backed access for modern services, a local or federated layer for systems that cannot integrate cleanly, and a privileged access path for administrative or emergency use. In practice, this means scoping passwordless and phishing-resistant MFA to the resources that can support them, then using compensating controls such as local break-glass accounts, remote access brokers, or dedicated admin workstations elsewhere.

For hybrid Windows estates, pay close attention to the sync boundary and the accounts that bridge it. A dedicated sync or connector account should be tightly scoped, monitored, and isolated from interactive use. For OT and legacy platforms, avoid collapsing operator access, maintenance access, and vendor access into one shared path, because that makes revocation and incident response much harder.

Where service or infrastructure accounts are involved, review them as identity objects, not just technical dependencies. AD and Entra ID hardening should be used to map privileged groups, delegation, service accounts, and hybrid identity paths before you decide which controls must stay local and which can be centralised.

For organisations trying to replace brittle secrets and passwords with stronger federation, one useful pattern is to move what can be modernised first, then narrow the exception set over time. Replacing app secrets with federated credentials is a good example of how to reduce standing secret exposure in parts of the environment that do support modern identity plumbing.

What Good Looks Like When Coverage Is Uneven

Good design does not force a single identity technology onto every asset. Instead, it creates a controlled exception model: modern systems use the strongest available Entra controls, legacy systems have explicit local governance, and OT systems use a documented access path with compensating security. That gives you continuity without pretending all platforms can accept the same control set.

Logging should also reflect the layering. You need cloud sign-in visibility for Entra-backed access, but you also need local authentication records, remote access logs, and evidence of who used brokered paths into OT or on-premises systems. Without that, a blended identity model becomes a blind spot rather than a control strategy.

Where the environment includes plant, HMI, SCADA, or other industrial assets, the identity model should be tested against outage constraints as well as security requirements. The right control is often not “centralise everything”, but “centralise governance, then segment enforcement by system capability and operational tolerance.”

Risk and Threat Considerations

When organisations assume Entra ID alone defines the identity perimeter, the usual failure modes are account sprawl, overbroad trust between environments, and weak fallback access paths that are never reviewed. In OT and legacy environments, that can leave shared credentials, dormant admin paths, or vendor routes as the easiest way in.

Failure mechanism: An attacker or insider compromises the strongest modern layer, then pivots through poorly governed local accounts, sync accounts, or remote access exceptions that were never brought under the same review and revocation discipline.

Impact: The result can be lateral movement from modern cloud access into on-premises or OT systems, loss of revocation certainty, and slower containment because the organisation cannot tell which identity layer was actually used.

Standards & Framework Alignment

This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.

NIST SP 800-63 and NIST SP 800-53 Rev 5 set the governance and control requirements practitioners need to meet.

FrameworkControl / ReferenceRelevance
NIST SP 800-63Digital Identity GuidelinesPhishing-resistant auth and assurance levels are central to layering modern identity controls.
Recommendation — Apply digital identity assurance guidance to the systems that can support modern authentication.
NIST SP 800-53 Rev 5IA-2 — Identification and Authentication (Organizational Users)Modern user sign-in and MFA layering depend on strong organisational authentication controls.
IA-5 — Authenticator ManagementLegacy and hybrid estates still need lifecycle control over passwords, secrets, and fallback authenticators.
IA-9 — Identification and Authentication (Non-Organizational Users)Vendor and third-party access to legacy or OT systems needs separate authentication governance.
Recommendation — Enforce strong user authentication for systems that can federate to Entra ID. Manage fallback authenticators tightly and rotate or retire them on a defined schedule. Treat third-party and external access paths as distinct authentication flows with explicit control.

Practitioner Guidance

What to prioritise: Start by inventorying every identity path that is not native to Entra ID, including local admins, service accounts, vendor logons, and offline operator access. If you cannot revoke it quickly, it needs stronger monitoring and tighter scope.

What to verify: Confirm which systems truly support phishing-resistant authentication, which only support local auth, and which rely on an intermediary. Then verify that emergency or fallback access is separately controlled, not just reused from the normal admin path.

Decision rule: If a system cannot consume modern Entra controls, do not treat it as an exception with no structure, treat it as a separate identity domain with explicit ownership, logging, and periodic recertification.

Practitioner takeaway: The goal is not universal identity uniformity, it is consistent governance across different technical realities, with the weakest layer constrained by design rather than accepted by default.

Free weekly newsletter

Subscribe to the NHI & AI Identity Journal

The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.

Bonus 33% off our NHI Course when you subscribe.

NHIMG Editorial Note
Reviewed and updated by the NHIMG editorial team on October 30, 2026.
NHI Mgmt Group — the #1 independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org