Right-sizing agent and service account access for least privilege

At a glance

What this is: This RSAC 2026 video explains why service accounts and AI agents need different access controls than human users, with least privilege and ownership as the central themes.

Why it matters: It matters because IAM teams cannot govern non-human identities safely with human-centric PAM and review cycles when access is broad, fast-moving and often tied to no clear owner.

👉 Watch P0 Security's RSAC 2026 video on right-sizing agent and service account access

Context

Non-human identities are identities that act through software rather than people, including service accounts, workloads and AI agents. The governance problem is not simply that they exist at scale, but that their access patterns are different from human logins and make traditional review and approval models weaker.

This article is about the access gap that opens when organisations give non-human identities broad permissions and treat them like ordinary users. For IAM, PAM and NHI programmes, the core issue is whether access is owned, scoped, rotated and revoked in a way that still works when the actor is a machine or an agent.

Key questions

Q: What breaks when non-human identities have more access than they need?

A: When non-human identities carry excess access, a single compromise can move from a local incident to broad cloud control. Over-privileged service accounts, tokens, and AI agents can reach storage, compute, and IAM functions that were never necessary for their job. The result is larger blast radius, faster lateral movement, and much harder containment.

Q: Why do service accounts and AI agents need different controls from human users?

A: Service accounts and AI agents authenticate and act without the predictable patterns that human identity systems expect. They can operate across runtimes, scale quickly, and carry permissions into automated workflows. That means access decisions should consider workload context, runtime behaviour, and time-bound authority rather than relying only on user-centric IAM patterns.

Q: How can organisations tell whether NHI governance is actually working?

A: NHI governance is working when every machine identity has an owner, a purpose, a minimum-necessary entitlement, and evidence of rotation and review. If teams can produce that chain without manual reconstruction, the programme is mature enough to withstand audit pressure. If they cannot, the governance model is still fragmented.

Q: When should organisations prioritise just-in-time access for AI agents over standing credentials?

A: Organisations should prioritise just-in-time access when AI agents need elevated permissions only for specific tasks, environments, or short windows of time. It is especially important for production systems, sensitive data paths, and cross-domain actions. JIT reduces standing exposure, limits reuse of stolen credentials, and makes approval and audit trails clearer.

Technical breakdown

Why service accounts and agentic identities need different access models

Service accounts usually follow predefined execution paths, while agentic identities can choose actions from a capability set to pursue a goal. That difference changes the control problem. Human-oriented PAM assumes a person logs in, completes a task and logs out, but non-human identities can act repeatedly, at machine speed and across multiple systems. If access is granted broadly at creation time, the control boundary becomes the initial provisioning decision rather than the current task context. Practical implication: model service accounts and agents separately, because one lifecycle and one entitlement pattern will not fit both.

Practical implication: define different governance rules for predefined service identities and goal-driven agentic identities.

How ownership, entitlements and lifecycle management interact

The article separates credential lifecycle from entitlement lifecycle, and that distinction matters. Credentials answer whether the identity can authenticate, while entitlements determine what it can do once authenticated. If ownership is unclear, neither lifecycle can be managed cleanly because no accountable human or team can approve change, rotation or revocation. That creates an auditability gap, not just an access gap. The technical problem is not merely whether access exists, but whether the organisation can trace it back to a responsible owner and a current business need. Practical implication: inventory both credentials and permissions together, then map each non-human identity to a named owner.

Practical implication: treat ownership as a control dependency for both credential and entitlement lifecycle management.

Why runtime authorization matters more than after-the-fact detection

Runtime authorization means evaluating access in context at the moment of action, rather than discovering misuse later. That is increasingly important for agents and service accounts because they can touch sensitive tools and data continuously, and their useful behaviour can look normal until the scope is crossed. Detection still matters, but it is no longer the primary boundary control when the identity is active throughout the task. Least privilege and just-in-time access reduce the standing blast radius before the action happens, while runtime checks enforce that the action still fits the current context. Practical implication: move high-risk authorisation decisions to execution time, not post-incident review.

Practical implication: push high-risk access decisions to runtime and reduce standing privilege before execution begins.

Breaches seen in the wild

• Cisco Active Directory credentials leak 2025: Kraken leaked Cisco Active Directory hashes, including service and krbtgt accounts; Cisco says they came from its 2022 breach, not a new one.
• Azure Key Vault Contributor escalation 2024: Datadog found Azure Key Vault Contributor could add itself to access policies and read every secret, key and certificate in a vault.

Read our 52 NHI Breaches Analysis report for a comprehensive view of breaches impacting Non-Human Identities including AI Agents.

NHI Mgmt Group analysis

Standing access is the wrong default for non-human identities: the article correctly frames broad, persistent permissions as the root governance failure. Service accounts and agents do not need human-style convenience access, because their work patterns are machine-paced and often repetitive. The practitioner conclusion is that standing privilege should be treated as an exception requiring explicit justification, not as the setup norm.

Ownership is the control that makes NHI governance enforceable: without a named owner, access reviews, revocation decisions and audit trails lose operational meaning. This is especially true for AI agents, where capability-based behaviour can expand faster than teams can document it. The implication is that account records must be tied to accountable teams before organisations can claim control over the identity estate.

Just-enough privilege is more defensible than broad agent access: the article points to a real shift from after-the-fact detection toward preventive controls. That direction aligns with OWASP-NHI and zero-standing-privilege thinking because machine identities do not need the same broad role assumptions as human users. Practitioners should treat scope as dynamic, but the governance rule remains static: grant only what the task needs.

Runtime authorisation gap: the article exposes the weakness of relying on periodic review for identities that act continuously. Access decisions made at provisioning time quickly drift from operational reality when agents and service accounts operate across changing contexts. The practical conclusion is that authorisation must be evaluated at execution time if the organisation wants meaningful control over machine action.

The next NHI maturity step is lifecycle coherence, not more inventory alone: visibility is necessary, but it is not sufficient if credentials, permissions and ownership are managed in separate workflows. The field is moving toward coordinated governance across issuance, rotation, revocation and entitlement change. Practitioners should use inventory as the starting point, then close the loop across the full NHI lifecycle.

From our research library:

• Only 5.7% of organisations have full visibility into their service accounts, according to the Ultimate Guide to NHIs.
• Read next: Service Account Security Guide

What this signals

Runtime authorisation is becoming the control boundary that matters most: access reviews and periodic recertification were built for identities whose privileges persist long enough to be observed. Agents and service accounts can complete work inside a much shorter window, so governance has to move closer to issuance and execution.

Ownership is now an access-control requirement, not just an administrative detail: when no accountable human or team is tied to the identity, access changes become ambiguous and revocation slows down. That is how NHI sprawl turns into audit failure, especially in environments where the identity estate already lacks visibility.

Just-enough privilege should be the default for machine action: the enterprise should not assume utility requires broad access. The control question is whether each identity can be constrained to the exact task context without losing operational function.

For practitioners

• Split service account and agent governance models Define separate policies for predefined service identities and goal-driven AI agents, because their access patterns, escalation paths and review cadence are not the same.
• Assign a human owner to every non-human identity Require a named accountable team or person for each service account and agent so access changes, revocation and audit responses have a clear decision-maker.
• Move high-risk permissions to just-in-time issuance Replace broad standing permissions with task-scoped access for sensitive tools and data, especially where agents can act across multiple systems.
• Separate credential lifecycle from entitlement lifecycle Track issuance, rotation and revocation alongside permission grants, changes and removals so one workflow does not mask failures in the other.
• Enforce runtime authorisation for agent actions Evaluate access in context at execution time for agents and service accounts that can reach sensitive resources continuously.

Key takeaways

• Non-human identities fail under human-centric access assumptions when organisations give them standing permissions that outlive the task they were meant to perform.
• The key governance gap is not just visibility, but accountability, entitlement scope and runtime control over what the identity can do right now.
• Programs that move toward just-enough privilege and task-scoped authorization reduce blast radius more effectively than post-fact detection alone.

Key terms

• Non-Human Identity (NHI): A digital identity assigned to a non-human entity such as a software application, service account, API key, bot, machine, or AI agent that enables it to authenticate and interact with systems without direct human involvement. NHIs now outnumber human identities in most enterprises by 25 to 50 times.
• Standing Privilege: Standing privilege is access that remains active even when no immediate task requires it. For NHI programmes, it is a common failure mode because long-lived credentials and persistent roles create unnecessary exposure. Reducing standing privilege usually means tighter expiry, on-demand access, and clearer review of who or what still needs access.
• Runtime Authorisation: Runtime authorisation is the practice of deciding access while a task is in progress, rather than only at provisioning time. It matters for NHIs because credentials and entitlements can change risk mid-session, especially when automation or AI agents interact with sensitive systems.
• Entitlement Lifecycle: The entitlement lifecycle covers how access is created, reviewed, used, changed, and removed over time. Strong lifecycle control prevents old permissions from lingering after a role, project, or need has ended, which is essential for least privilege and audit readiness.

What's in the full article

P0 Security's full video covers the operational detail this post intentionally leaves for the source:

• How P0 Security frames right-sizing workflows for service accounts and AI agents in production
• The article's practical distinction between credential lifecycle and entitlement lifecycle management
• Examples of runtime authorization for SOC and HR agent use cases
• The source discussion of how organisations can shift from detection-first to preventive controls

👉 The full P0 Security video covers ownership, runtime authorization and least-privilege controls for non-human identities and AI agents.

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 building or maturing an IAM programme, it is worth exploring.