Extended access management for AI agents and unmanaged identities

At a glance

What this is: This is a product and category explanation from 1Password arguing that traditional IAM, IGA, and MDM leave an access-trust gap across unmanaged apps, devices, unfederated identities, and AI agents.

Why it matters: IAM, NHI, and governance teams need to understand where access controls stop, because unmanaged and non-federated identity paths now sit outside the controls many programmes assume are universal.

Context

Extended access management is 1Password’s term for closing the gap between who or what can sign in and what legacy IAM, IGA, and MDM tools can actually govern. In practical terms, the problem is not only user identity on managed endpoints, but every sign-in path that still reaches sensitive data without full lifecycle visibility.

The article ties that gap to four conditions: unfederated identities, unmanaged apps, unmanaged devices, and AI agents. For identity teams, that combination matters because governance models built around federation, device management, and static entitlement review do not fully describe the access surface anymore.

Key questions

Q: How should teams govern access when unmanaged apps and devices still reach sensitive data?

A: Teams should govern by access path, not just by account status. The practical test is whether the organisation can inventory, restrict, and revoke access when the device, app, or identity sits outside normal managed controls. If it cannot, that access should be treated as outside trusted governance even if authentication succeeds.

Q: Why do AI agents complicate existing IAM and NHI governance models?

A: AI agents complicate governance because access is no longer confined to a single environment or a single identity type. An agent may need cloud runtime permissions, customer data access, and tool-level OAuth tokens at the same time, which means standing privilege and lifecycle assumptions break down fast. That is why one control model rarely covers the full path.

Q: What breaks when federated identity is the only control used for access governance?

A: Federation covers a narrow part of the problem. It does not guarantee that the device is managed, the app is inventoried, or the actor is operating inside a full governance boundary. As a result, teams can end up with authenticated sessions that are still outside meaningful access control.

Q: When should security teams treat access as unmanaged even if it is authenticated?

A: Treat access as unmanaged whenever the organisation cannot prove ownership of the endpoint, control of the application, or lifecycle visibility over the identity involved. That is the point at which authentication no longer equals governance, and the access path should be reviewed as a separate risk surface.

Technical breakdown

Why the access-trust gap appears outside legacy IAM coverage

Legacy IAM assumes sign-in happens through managed, federated, and policy-visible paths. When an identity uses an unmanaged app or device, or when an AI agent operates against data sources outside the normal control plane, the control relationship changes: the system can still authenticate access, but governance no longer has the same inventory, policy, or revocation reach. That is why the article frames the issue as an access-trust gap rather than a simple access-management problem. The important distinction is visibility versus assurance. Security teams may know a sign-in occurred, but not whether the path, device state, or identity type was actually governed end to end. Practical implication: map where your current IAM stack loses lifecycle control, not just where it still authenticates users.

Practical implication: map where your current IAM stack loses lifecycle control, not just where it still authenticates users.

How unmanaged devices and apps change identity assurance

An unmanaged device or app can participate in access without being inside the enterprise’s normal posture, inventory, or enforcement model. That breaks the assumption that identity assurance can be derived from central control of the endpoint or application estate. For NHI and human IAM teams alike, the risk is that access decisions look complete at the authentication layer while the device, app, or client remains outside the governance boundary. In that situation, entitlement review alone does not tell you whether the path to data is actually trustworthy. The operational question becomes whether the programme can detect and govern access from assets it does not administer directly. Practical implication: classify unmanaged endpoints and apps as governance blind spots, not edge cases.

Practical implication: classify unmanaged endpoints and apps as governance blind spots, not edge cases.

What AI agents add to non-human identity governance

AI agents change the problem because they are not just tools that consume credentials; they can participate in access decisions and runtime actions across multiple systems. That makes them a non-human identity governance issue, not only an application or automation issue. If an agent can interact with sensitive data, call tools, and traverse services, then its identity, privilege scope, and offboarding path all matter. The article’s framing is important because it places AI agents in the same governance conversation as unmanaged apps and devices: access is only trustworthy if the actor, its credentials, and its permitted actions are all governed together. Practical implication: treat AI agents as identities with lifecycle, privilege, and audit requirements, not as generic automation.

Practical implication: treat AI agents as identities with lifecycle, privilege, and audit requirements, not as generic automation.

NHI Mgmt Group analysis

Access assurance is now a path problem, not just an identity problem: The article correctly shows that legacy IAM, IGA, and MDM do not fully describe the access surface when unmanaged apps, devices, and unfederated identities remain in use. The field has treated authentication and inventory as sufficient proxies for governance, but those controls fail when the sign-in path itself is outside administrative reach. Practitioners should now measure governance by path coverage, not by login coverage alone.

The Access-Trust Gap is a useful named concept because it captures a structural failure in current identity programmes: the organisation trusts access decisions that were never made with complete context. That gap is especially visible when AI agents and unmanaged endpoints participate in data access without the same lifecycle controls as managed users. The implication is that identity governance has to expand from account-centric oversight to actor-and-path assurance.

AI agents belong in NHI governance because their access behaviour is operational, persistent, and policy-relevant: even when they are not fully autonomous, they can still sit in a non-human identity lifecycle that includes provisioning, privilege scope, monitoring, and offboarding. Treating them as mere tooling obscures the governance question: who or what is authorised to act, on which systems, under which conditions? The practical conclusion is that NHI programme boundaries now need to include AI agents wherever they touch enterprise data.

Unmanaged devices and apps expose the limits of federated thinking: federation solves part of authentication, but it does not by itself govern every endpoint, every application, or every runtime context. This article highlights a broader market direction toward control-plane coverage over all sign-in paths, which will pressure IAM teams to reconcile device posture, application inventory, and identity lifecycle under one governance model. Practitioners should expect access governance to become broader, not narrower, as the estate becomes more heterogeneous.

Extended access management is best understood as governance over every sign-in path: that phrasing matters because it shifts the discussion away from products and toward control boundaries. The challenge for practitioners is not whether a platform can authenticate a session, but whether the programme can account for unmanaged actors, unmanaged clients, and agentic access with the same discipline applied to managed users. The next maturity step is end-to-end accountability for access paths, not just for identities.

What this signals

Access governance is moving from account-centric control to path-centric assurance: programmes that only track managed identities will miss the exposure created by unfederated users, unmanaged clients, and AI agents that can still reach sensitive data. The practical shift is to ask where access enters the estate, not just who holds the account.

Extended access management is a sign that identity boundaries are widening: IAM teams are now expected to reconcile human identity, NHI, and emerging agentic access under one governance model. That does not replace federation or device management, but it does mean those controls are no longer sufficient on their own.

For practitioners

• Map every sign-in path Inventory which apps, devices, identities, and AI agents can reach sensitive data without passing through managed federation, device posture, or normal lifecycle control.
• Classify unmanaged access as a governance gap Separate authenticated access from governed access in your programme reporting so unmanaged endpoints and apps are tracked as blind spots, not exceptions.
• Add AI agents to identity lifecycle scope Define provisioning, privilege review, logging, and offboarding for any AI agent that can access enterprise systems or data.
• Test revocation outside managed estates Validate whether you can actually remove access when the actor is an unfederated identity, a contractor device, or an unmanaged client.

Key takeaways

• The article argues that legacy identity controls leave an access-trust gap wherever identities, apps, or devices sit outside managed governance.
• AI agents belong in the same lifecycle conversation as other non-human identities when they can reach enterprise data and tools.
• Practitioners should focus on access path coverage, revocation certainty, and lifecycle visibility rather than relying on authentication alone.

Key terms

• Access-trust gap: The gap between having a policy and actually enforcing it when access is requested. It appears when compliance, HR, or training data sits in separate systems from the access decision, allowing users to reach resources even though a required condition has not been met.
• Extended access management: Extended access management is a control model that covers access beyond traditional employee login, including SaaS, service identities, and AI agents. It tries to unify trust, entitlement, and device context so organisations can govern access wherever work actually happens.
• Unmanaged identity: Any identity that can authenticate or act without being fully controlled by the organisation’s standard identity stack. That includes service accounts, API keys, tokens, and some AI agents. The risk is not just visibility loss. It is loss of ownership, lifecycle control, and reliable revocation.
• Agentic Access: Agentic access is delegated system access granted to an AI agent or autonomous workflow so it can perform defined tasks across tools and data sources. It differs from human access because the actor can execute continuously, combine actions quickly, and amplify mistakes at scale.

Deepen your knowledge

NHI governance, agentic AI identity, and machine identity lifecycle are core topics in our NHI Foundation Level course, the industry's only accredited NHI security programme. If you are responsible for identity security strategy or NHI governance in your organisation, it is worth exploring.