Token Security’s NHI governance model for cloud and AI sprawl

At a glance

What this is: This is a Token Security analysis of why non-human identity governance breaks down across cloud, containers, and agentic AI, with scale and fragmentation as the central finding.

Why it matters: It matters because IAM, PAM, and NHI programmes now have to govern identities that are created, used, and retired across multiple systems faster than traditional review and remediation cycles can track.

By the numbers:

• Non-human identities now represent more than 98% of all identities within the enterprise.

Context

Non-human identity governance breaks down when identities are created, used, and discarded across cloud platforms, container platforms, and agentic AI systems faster than teams can review them. The primary problem is not just volume, but fragmentation: different clouds, tools, and teams hold partial context for the same identity.

In that environment, service accounts, workload identities, access tokens, APIs, and autonomous agents behave more like a distributed control surface than a static inventory. Traditional IAM models were built around centralized human administration, so they struggle when remediation depends on correlating runtime usage, ownership, and downstream dependency before action is taken.

Token Security's article frames this as a modern NHI governance problem for cloud-native and AI-enabled enterprises. That framing is typical of organisations moving from visibility gaps to lifecycle and remediation gaps at the same time.

Key questions

Q: What breaks when NHI governance is still built around static identity records?

A: Static identity records break down when accounts, tokens, and service principals are created and consumed across multiple clouds and runtime systems faster than teams can certify them. The control failure is not just lack of inventory. It is lack of correlated context for ownership, runtime use, and dependency, which makes safe remediation impossible.

Q: Why do multi-cloud AI environments increase NHI risk?

A: They increase NHI risk because service accounts, tokens, and model endpoints are split across different IAM systems with incompatible policy models. That creates more places for over-permission, missed offboarding, and hidden data flows to accumulate before anyone notices.

Q: How should security teams decide whether an NHI is safe to remediate?

A: Security teams should require evidence, not intuition. A remediation decision should include ownership, active dependency signals, usage patterns, and the expected blast radius if access changes. When those inputs are missing, the finding should stay in investigation rather than being forced into action. That is the difference between visibility and enforceability.

Q: How should security teams govern AI agents that use existing NHI credentials?

A: Treat every AI agent as a non-human identity with its own owner, scope, and expiry. Assign the minimum permissions needed, issue short-lived credentials where possible, and log every action the agent takes. Governance should focus on both access and behavior, because a valid credential can still be used in an unsafe way.

Technical breakdown

Why fragmented cloud identity models break NHI governance

Hybrid and multi-cloud environments create separate identity silos, each with its own permission model, credential format, and ownership boundary. AWS IAM roles, Azure service principals, GCP service accounts, SaaS OAuth tokens, and on-prem service accounts do not resolve into one control plane automatically. The technical problem is not simply inventory size, but the lack of normalized context across systems that were not designed to share lifecycle state. Without that context, teams cannot confidently answer who issued the identity, where it is used, or what dependency it supports.

Practical implication: build a unified identity inventory before attempting centralized remediation or policy automation.

How ephemeral workloads amplify overprivilege and secret sprawl

Kubernetes, Docker, serverless functions, and CI/CD pipelines generate identities that can exist for seconds and still inherit privileges, tokens, or secrets. Because these identities are short-lived, they often bypass the governance rhythms built for slower human lifecycle events. That makes over-provisioning and orphaned access more likely, especially when creation is developer-led and not mediated through a consistent control process. In effect, the credential lifetime and the governance lifetime no longer match.

Practical implication: align credential issuance, runtime use, and revocation controls to short-lived workload behaviour rather than human review cadences.

Why agentic AI changes the identity control problem

Agentic AI systems do not just consume credentials; they initiate actions, make API calls, and operate across systems on their own runtime timeline. That means the identity is no longer a passive subject waiting for a request, but an active actor choosing when to authenticate and where to act. When those agents authenticate with unmanaged credentials, the trust boundary shifts from fixed ownership to dynamic behavior. The technical challenge becomes correlating access, data use, and delegation across sessions and systems, not just assigning an account record.

Practical implication: treat agent access as runtime behavior to govern, not as a static entitlement to review later.

NHI Mgmt Group analysis

Identity as a data problem is now the correct operating model for NHI governance: fragmented cloud and AI environments produce too many identity events for manual or point-in-time control to scale. The important shift is not only volume, but correlation, because remediation decisions depend on usage, ownership, and dependency context that legacy tools keep separate. Practitioners should treat identity telemetry as a governed dataset, not a ticket queue.

More than 98% identity concentration changes the control equation: when non-human identities dominate the estate, even small governance gaps create disproportionate exposure. That scale means policy drift, stale secrets, and unowned accounts are no longer edge cases, they are the baseline failure mode. The implication is that identity programmes must be designed around machine and workload behaviour first, not adapted from human-centric patterns.

Big data identity correlation is the missing bridge between visibility and safe remediation: teams do not fail because they cannot see that an identity exists, they fail because they cannot determine what will break if they act on it. Runtime analytics, ownership mapping, and dependency tracing are what make deprovisioning or narrowing access safe at enterprise scale. Practitioners should define remediation confidence as a governance requirement, not a nice-to-have.

Autonomous behaviour collapses the assumption that access is stable long enough to review: access review processes were designed for identities whose privileges persist across a certification cycle. That assumption fails when an agent can acquire and use credentials within a single task while initiating its own API calls. The implication is not a better review cadence, but a different governance model for actors that create and consume access inside the same runtime window.

Shadow AI discovery belongs inside NHI governance, not beside it: custom GPTs and other agentic systems using unmanaged credentials are not a separate problem class, they are a new path into the same identity estate. The practical consequence is that NHI inventory, secrets management, and AI oversight must converge around the same ownership and usage graph. Practitioners should govern agent identity as part of the core NHI programme, not as an adjacent AI project.

From our research library:

• 67% of organisations still rely heavily on static credentials despite the risks they pose to agentic AI deployments, according to the 2026 Infrastructure Identity Survey.
• Only 13% of organisations feel extremely prepared for the reality of agentic AI despite the majority racing toward autonomous adoption, according to the 2026 Infrastructure Identity Survey.
• Read next: Ultimate Guide to NHIs

What this signals

Identity operations need a data-layer reset: when non-human identities span clouds, containers, CI/CD, SaaS, and AI systems, the real control problem becomes correlation across sources, not another point solution. Teams that cannot normalize ownership and runtime context will keep remediating symptoms instead of governance state.

Autonomous agents force a timing change in governance: certification cycles assume access persists long enough to be observed and reviewed. That assumption weakens when an agent acquires credentials, acts, and exits inside one workflow, which pushes governance closer to issuance time and runtime policy enforcement.

Shadow AI discovery should sit inside the NHI programme: unmanaged custom GPTs and agentic systems are just another route into the identity estate, but with faster action cycles and weaker human accountability. The practical signal for practitioners is that secrets management, AI oversight, and NHI inventory now need one shared operating picture.

For practitioners

• Build a unified NHI inventory Correlate identities across AWS, Azure, GCP, Kubernetes, SaaS, CI/CD, secrets managers, and on-prem systems so ownership and runtime use are visible in one place.
• Map runtime usage before remediation Trace how each credential, token, or service account is actually consumed before revoking or narrowing access, especially where production dependencies are opaque.
• Separate short-lived workloads from human review cadences Design governance around ephemeral infrastructure by tying issuance, use, and revocation to workload duration rather than periodic certification cycles.
• Inventory shadow AI and agentic access paths Identify custom GPTs, autonomous agents, and other AI systems that authenticate with unmanaged credentials, then fold them into the same ownership and policy model as other NHIs.
• Automate policy with dependency awareness Use contextual dependency mapping so auto-deprovisioning, key rotation, and access narrowing happen only when downstream impact is understood.

Key takeaways

• Non-human identity risk in modern enterprises is driven by fragmentation as much as scale, because cloud, container, and AI systems each create their own control gaps.
• The article states that NHIs now represent more than 98% of enterprise identities, which means machine and agent governance is now the dominant identity problem.
• Safe remediation depends on correlating ownership, runtime use, and downstream dependency before policy is enforced or access is removed.

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.
• Identity correlation: Identity correlation is the process of linking multiple account records to one governed subject. It lets IAM and IGA teams understand that separate usernames, principals, or emails may belong to the same employee or workload, which is essential for access review, offboarding, and entitlement analysis.
• Runtime Dependency Risk: Runtime dependency risk is the gap between what a software manifest says will run and what a live workload is actually executing. It matters because malicious or altered packages can behave differently at runtime, especially when they touch secrets, spawn processes, or call external endpoints.
• Shadow AI: AI agents, copilots, or connected tools operating without full visibility or governance from security teams. Shadow AI becomes an identity problem when those systems authenticate with unmanaged tokens, service accounts, or OAuth apps that can reach production resources.

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.