AI agent governance and identity standards take centre stage on 5 November 2026

At a glance

What this is: NHI Mgmt Group's analysis, drawing on Opal Security's event briefing: Security and identity leaders will discuss how to govern AI agents with ownership, access reviews, just-in-time access and emerging identity standards.

Why it matters: IAM, IGA and PAM teams need to decide whether their current controls can govern agent identities that act at machine speed and may outlive human review cycles.

By the numbers:

• On 5 November 2026, the briefing runs from 10am to 12pm PT as a virtual event.

👉 Register for Opal Security's virtual briefing on AI agent governance and identity standards

Context

AI agent governance is moving from theory to operating model. The problem is no longer whether agents can act, but whether existing identity programmes can assign ownership, constrain access, and later explain what an agent did. That becomes more difficult when the same system can create, combine, and use privileges during runtime rather than only through pre-approved workflows.

This conversation sits at the boundary of IAM, IGA, PAM, and emerging agent identity standards. The practical question is whether teams can extend review, attribution, and least-privilege controls to non-human actors without relying on assumptions that were built for people.

Key questions

Q: How should security teams govern AI agents that inherit authority from other identities?

A: Security teams should govern AI agents by tracking identity lineage, not just credentials. That means recording the originating identity, the delegated authority path, and the runtime context for each action. If an agent can inherit permissions from humans, services, or other agents, policy has to evaluate the full chain before access is granted or continued.

Q: When does just-in-time access help most for AI agents?

A: JIT access helps most when agent tasks are episodic, high-risk, or difficult to predict in advance. It reduces standing exposure, but it works only if the agent's permissions are also tightly scoped and actively revoked after use. Otherwise, the temporary token masks a persistent privilege problem.

Q: What are the warning signs that an AI agent's access model is failing?

A: The warning signs are persistent permissions, unclear ownership, poor attribution and actions that cannot be tied back to a specific policy decision. If a team cannot explain who approved the access, what the agent was allowed to do, and how the action will be audited, governance is already failing.

Q: Why do machine identities matter in IAM programmes?

A: Machine identities matter because servers, applications, mobile devices, and IoT endpoints authenticate, exchange data, and often hold privileged access. If they are not inventoried and governed, security teams lose visibility into what is actually trusted in the environment. That creates blind spots in certificate lifecycle management and weakens the entire identity fabric.

Background and context

Why agent identity breaks human-paced access models

Traditional IAM assumes a person or service can be enrolled, reviewed, and offboarded through discrete lifecycle steps. AI agents complicate that model because they may request access, use tools, and complete actions in a short runtime window, often faster than governance cycles can observe. Identity for an agent is therefore not just authentication. It also includes ownership, delegated authority, scope, and evidence that the actor can be traced after the fact. The control problem shifts from static assignment to runtime governance, where the question is whether access decisions can follow the agent's actual behaviour rather than its nominal role. See the Agentic AI Identity Guide.

Practical implication: Treat agent identity as a governed runtime relationship, not just a registration record.

Just-in-time access versus standing access for AI agents

Standing access gives an agent persistent permissions that remain available whether or not the task justifies them. Just-in-time access narrows that window by issuing permissions only when needed and revoking them when the task ends. For AI agents, the distinction matters because the wrong pattern changes blast radius immediately: a broad standing grant turns an execution issue into persistent identity risk, while task-scoped access ties authority to the exact work being performed. The real design question is not whether automation is convenient, but whether access can be made ephemeral enough to match the agent's operating tempo. See the AI Agent Authorisation Guide.

Practical implication: Prefer task-scoped issuance and revocation over durable agent entitlements.

How emerging identity standards make agent actions auditable

Agent identity standards aim to let policy systems decide what an agent may do and later prove what it actually did. That requires consistent identity signals across registration, delegation, authorization, and logging. Without those signals, teams may know that an agent existed but not which tools it used, under whose authority it operated, or whether a policy decision was enforced in practice. Standards work matters because it turns agent behaviour from an opaque runtime event into something security controls can evaluate, log, and correlate. In other words, identity standards are the bridge between agent autonomy and auditability. See the Agent Identity Standards Tracker.

Practical implication: Anchor agent controls to standards that preserve identity context across policy, tooling and logs.

NHI Mgmt Group analysis

Agent identity is no longer a theoretical extension of IAM, it is a governance category in its own right. Once an AI system can act, select tools, and time its own execution, the programme must govern ownership, authority and traceability as first-class identity properties. That moves the control discussion beyond authentication alone and into delegated responsibility across the full lifecycle. The practical conclusion is that agent identity now belongs in the same operating model as human and machine identities.

Access review processes assume privilege persists long enough to be reviewed, but agentic access often does not. That assumption was designed for actors whose permissions remain stable between certification cycles. It fails when an agent acquires, uses and releases access within a short runtime window, because there is no durable state to inspect after the fact. Audit and incident response becomes more important, but the deeper point is that the review model itself was built for slower actors.

Standing access is becoming the wrong default for AI agents. The event's emphasis on just-in-time access reflects a broader shift in identity design: agent permissions should match task duration, not organisational comfort. Persistent grants turn every successful agent interaction into future blast radius. Practitioners should read this as a signal that entitlement design is moving from convenience-first provisioning to time-bounded authority.

Emerging agent identity standards matter because auditability is now a prerequisite for trust. The field cannot rely on generic logs or application-level traces to explain agent behaviour after a security event. Standards that preserve delegation, policy and identity context are what make accountability durable enough for governance and incident review. Security teams should expect standards maturity to shape how quickly agent identity becomes operationally governable.

Runtime identity governance is the named concept this market is converging on. It captures the shift from reviewing identities after access has been granted to governing authority while an agent is still executing. That is a different control problem from traditional IAM, and it explains why leaders are combining ownership, reviews, least privilege and standards into one conversation. The implication is that agent identity programmes will be judged on runtime enforcement, not policy intent alone.

From our research library:

• 19% of organisations give AI systems dramatically more access than human employees, nearly one in five granting unrestricted privilege, according to the 2026 Infrastructure Identity Survey.
• 70% of organisations grant AI systems more access than they would give a human employee performing the exact same job, according to the 2026 Infrastructure Identity Survey.
• Read next: AI Agent Authorisation Guide

What this signals

Agent identity governance will increasingly be judged by runtime decisions, not by whether a policy exists on paper. The more an organisation lets agents act with durable privileges, the more it needs proof that those privileges were bounded in time and scope. That shifts programme emphasis toward issuance-time controls, attribution and revocation, not just periodic review.

Access review cadences will need to shrink around autonomous execution windows. Governance models that assume a stable entitlement state between review cycles will miss the most important events. The operational question becomes whether access can be observed, evaluated and withdrawn fast enough to match agent speed.

Emerging standards will separate teams that can audit agent behaviour from those that can only describe it. When identity, delegation and tooling context are preserved, security leaders can make accountability real instead of approximate. The practical signal is that agent identity standards are becoming a control plane issue, not just an interoperability topic.

For practitioners

• Inventory every production agent Build a complete register of agents, their owners, their delegated purposes and the systems they can reach.
• Replace standing grants with task-scoped access Constrain permissions to the minimum needed for each agent task and revoke them when the task completes.
• Attach traceability to each agent action Log the actor, the delegated authority, the tool used and the policy decision so you can reconstruct what happened later.
• Set a review cadence for agent entitlements Treat agent access as a living entitlement set that requires periodic attestation, not a one-time onboarding event.

Key takeaways

• AI agents are forcing identity teams to extend ownership, scope and accountability beyond human-centric IAM assumptions.
• Persistent access is the clearest governance risk because it lets agent behaviour outlive the task that justified it.
• The strongest control pattern is not more review after the fact, but tighter issuance, traceability and standards-backed governance at runtime.

Related NHI Mgmt Group resources: AI Agent Authorisation Guide · AI Agent Observability, Audit and Incident Response Guide · Agent Identity Standards Tracker · Agentic AI Identity Guide

Key terms

• Agent Identity: An agent identity is the set of attributes, credentials and permissions assigned to an autonomous software entity. It is treated as a non-human identity because it can authenticate, act on systems and accumulate access over time, which creates governance, audit and lifecycle obligations similar to other production identities.
• Delegated authority: Delegated authority is permission given to one identity to act on behalf of another within defined limits. In identity systems, it enables an agent, user, or service account to perform approved actions without owning the underlying rights, using scoped tokens, roles, or approvals to constrain access and accountability.
• Just-in-Time (JIT) Access: Just-in-Time (JIT) Access grants permissions only when needed and removes them after use. In IAM and PAM, it reduces standing privilege by issuing time-bound access based on approval, context, or policy, limiting exposure for humans, service accounts, and non-human identities while improving auditability and control.
• Access Review: A formal process for confirming whether access is still needed and justified. In IAM programs, the review becomes an evidence-bearing control when decisions are recorded, scoped correctly, and traceable to the right reviewer, application owner, or auditor.

What to expect at the briefing

Opal Security's full event briefing covers the operational detail this post intentionally leaves for the source:

• Live perspectives from security and identity leaders on ownership models for AI agents
• Discussion of the tradeoffs between standing access, just-in-time access and policy enforcement at agent speed
• Practical examples of how teams are extending access reviews to non-human identities
• Emerging standards questions that practitioners still need to resolve before agent identity can scale

👉 The full Opal Security event covers ownership models, access review practices and emerging agent identity standards.

Deepen your knowledge

NHI governance, agentic AI identity, and machine identity security are core topics in our NHI Foundation Level course, the industry's only accredited NHI security programme. If you are building or maturing an IAM programme, it is worth exploring.