Zero trust for non-human workloads still leaves NHI gaps

At a glance

What this is: This is a product-oriented analysis of zero trust for non-human workloads, with the key finding that workload access still depends on continuous identity verification, policy enforcement, and better credential management.

Why it matters: For IAM and NHI practitioners, the relevance is that workload access cannot be governed safely with user-era assumptions, especially when secrets, posture, and conditional access all need to be evaluated at runtime.

By the numbers:

• 97% of NHIs carry excessive privileges, increasing unauthorised access and broadening the attack surface.
• 61% rely on spreadsheets or manual tracking for machine identity management.
• 80% of identity breaches involved compromised non-human identities such as service accounts and API keys.

👉 Read Aembit's analysis of zero trust for non-human workloads and workload IAM

Context

Zero trust for non-human workloads means every application, service, or automation process must be authenticated and authorised before it can reach another system. That matters because machine access now often moves faster than human access, but governance, revocation, and review still lag behind. The primary keyword here is zero trust for non-human workloads, and the real question is whether identity teams can apply it without creating a new layer of unmanaged secrets.

The article reflects a common enterprise tension: teams want MFA-like controls and conditional access for workloads, yet most existing IAM programs were built around people, not autonomous processes. The model is directionally right, but the operational burden shifts to inventory, policy definition, and lifecycle control. For a deeper baseline on NHI scope and lifecycle risks, see the Ultimate Guide to NHIs and the Ultimate Guide to NHIs -- Key Challenges and Risks.

The posture described here is not unusual. Most organisations still struggle to maintain complete visibility into machine identities, which makes any promise of continuous verification only as strong as the underlying inventory and revocation process.

Key questions

Q: How should security teams apply zero trust to non-human workloads?

A: Security teams should treat each workload request as untrusted until identity, policy, and context are verified at runtime. That means short-lived credentials, explicit authorisation rules, and continuous revocation paths for service accounts, tokens, and certificates. Zero trust only works for workloads when the identity layer is complete and the control plane can enforce decisions without manual intervention.

Q: What is the difference between workload zero trust and traditional network segmentation?

A: Network segmentation limits where traffic can move, but workload zero trust decides whether a specific request should be allowed at all. Segmentation is about path control. Workload zero trust is about identity-based authorisation at the moment of access, which is far more useful when attackers can reuse machine credentials inside trusted environments.

Q: When do ephemeral NHI credentials create less risk than static secrets?

A: Ephemeral credentials reduce risk when the workload can reauthenticate automatically and the organisation can revoke access quickly. They are most effective for high-value services with clear ownership, strong inventory, and short operational windows. If the environment cannot maintain those controls, ephemeral credentials may simply hide the same governance gaps behind shorter lifetimes.

Q: Why do non-human identities complicate zero trust architecture?

A: Non-human identities complicate zero trust because they often operate at machine speed, across many environments, and with more privileges than teams realise. The result is a governance problem, not just a technical one. Without complete inventory, ownership, and lifecycle controls, zero trust becomes an assumption rather than an enforceable model.

Technical breakdown

How zero trust changes workload identity verification

Traditional workload access often assumes that anything inside a network boundary is acceptable once a credential is presented. Zero trust changes that model by requiring every request to be evaluated against identity, policy, and context before access is granted. For non-human identities, that context can include workload identity, source environment, time, location, and posture signals. The important distinction is that trust is not inherited from network position or deployment location. It is recomputed at request time, which makes policy quality and identity inventory the real control plane.

Practical implication: Map workload access to explicit policies and reject any access path that depends on implicit network trust.

MFA-like controls for non-human identities

MFA for humans proves presence through multiple factors. For workloads, the equivalent is conditional authorisation, where a machine identity must satisfy policy gates before it receives access. That often means combining identity proof, runtime context, and short-lived credentials rather than reusing a static secret. The security value comes from reducing the usefulness of stolen credentials and constraining where and when a workload can act. But the control only works if the policy engine has accurate identity data and if the workload can be re-authenticated without manual intervention.

Practical implication: Use short-lived, policy-bound access for workload connections that can be revalidated continuously.

Credential management and secretless workload access

The article points to eliminating low-level credential handling, which is the right direction for NHI governance. Static secrets create an audit problem, an offboarding problem, and a blast-radius problem because they persist long after the workload or job that needed them has changed. Secretless patterns reduce that exposure by moving from stored credentials to ephemeral access decisions, but they do not eliminate governance. The organisation still needs lifecycle controls, ownership, rotation logic, and revocation pathways. Without those, secretless access can become another abstraction over the same accountability gap.

Practical implication: Treat secretless design as a control improvement, not as a substitute for lifecycle governance.

Threat narrative

Attacker objective: The attacker seeks durable access through workload channels that bypass human-focused controls and enable lateral movement.

1. Entry occurs when an attacker obtains a workload credential or abuses a permissive workload identity that was never tightly scoped.
2. Escalation follows when the compromised non-human identity can reach downstream services without additional reauthentication or conditional checks.
3. Impact comes from lateral movement and data access inside service-to-service paths that were assumed to be trusted.

Breaches seen in the wild

• Shai Hulud npm malware campaign — Shai Hulud campaign: npm malware exposed secrets on GitHub.
• Reviewdog GitHub Action supply chain attack — reviewdog/action-setup GitHub Action supply chain attack exposed secrets.

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

NHI Mgmt Group analysis

Zero trust for workloads is now an NHI governance problem, not just an architecture pattern. The article treats zero trust as a way to extend human security controls to machines, which is directionally correct but incomplete. Workloads do not authenticate like people, and policy engines cannot compensate for weak identity inventories or poor offboarding. Practitioners should treat workload zero trust as a lifecycle governance discipline, not a network overlay.

Ephemeral access only reduces risk when identity proof, policy, and revocation stay in sync. Short-lived credentials narrow the window of abuse, but they do not solve trust drift if the workload identity itself is stale or over-scoped. In practice, the control boundary shifts from credential storage to runtime authorisation. Teams need to decide whether their policy model is expressive enough to support that shift before they deploy it at scale.

Credential removal without ownership creates a hidden governance gap. The article emphasises automating away manual secret handling, which is valuable, but automation cannot assign accountability on its own. If security, platform, and development teams cannot name the owner of each workload identity, the organisation will still struggle with review and revocation. The practitioner conclusion is simple: automation should enforce ownership, not replace it.

Workload zero trust should be measured by blast-radius reduction, not by the number of controls added. More checks do not equal better governance if they are bolted onto unclear identity boundaries. The better test is whether a stolen workload credential can still move laterally, persist, or reach sensitive data. If the answer is yes, the zero trust model is only partial and needs stronger NHI lifecycle controls.

Secretless architecture is becoming a governance expectation, but only for environments that can support it operationally. Eliminating long-lived secrets is a sound objective, yet many estates still depend on brittle provisioning paths and incomplete inventory. That means the immediate priority is not a blanket redesign but identifying where ephemeral access can replace static credentials first. Practitioners should phase the transition by risk and asset criticality.

From our research:

• 57% of organisations lack a complete inventory of their machine identities, according to The Critical Gaps in Machine Identity Management report.
• Only 38% have automated certificate lifecycle management in place, which leaves many workload access paths tied to manual processes and slow revocation.
• For a broader governance baseline, review the Ultimate Guide to NHIs alongside the lifecycle controls that make continuous verification workable.

What this signals

Ephemeral access only becomes meaningful when teams can actually see every machine identity they are granting it to. With 57% of organisations lacking a complete machine identity inventory, according to The Critical Gaps in Machine Identity Management report, the governance problem is still visibility before enforcement. Practitioners should expect workload zero trust programmes to stall unless ownership, inventory, and review are treated as prerequisites, not afterthoughts.

Identity blast radius: the practical measure here is how far a compromised workload credential can move before it is detected or revoked. That is why teams should align workload access policy with NIST SP 800-207 Zero Trust Architecture rather than layering policy on top of implicit trust zones.

For practitioners

• Implement runtime policy checks for workload access Require each workload-to-workload request to pass identity, posture, and context evaluation before the request is authorised. Start with sensitive services and external-facing integrations where lateral movement would matter most.
• Inventory every non-human identity before tightening access Build a complete list of service accounts, tokens, certificates, and automation identities, then map each one to an owner and a business process. Tie this inventory to the Ultimate Guide to NHIs and review it on a fixed cadence.
• Move high-risk workloads to short-lived credentials Prioritise systems that currently rely on persistent secrets and replace them with ephemeral credentials wherever the application architecture allows it. Document revocation paths so the credentials can be cut off without manual cleanup.
• Define offboarding for machine identities Create a revocation workflow for every workload identity so decommissioned apps, retired jobs, and rotated integrations cannot continue to authenticate. Use the Ultimate Guide to NHIs -- Key Challenges and Risks as the baseline for lifecycle controls.

Key takeaways

• Workload zero trust is an NHI governance problem because machines also need inventory, ownership, and revocation, not just policy labels.
• Short-lived access reduces exposure, but it cannot compensate for stale identities or incomplete lifecycle controls.
• The practical test is blast-radius reduction: if a stolen workload credential can still move laterally, the control model is not finished.

Key terms

• Non-Human Identity: A non-human identity is any credentialed actor that is not a person, including service accounts, API keys, tokens, certificates, and autonomous agents. In practice, these identities often outnumber human accounts and require the same lifecycle governance, but at much higher scale and speed.
• Workload Identity: Workload identity is the mechanism that lets an application or service prove who it is when talking to another system. It replaces assumptions based on network location or shared secrets with explicit identity, making access decisions more precise and easier to revoke when the workload changes.
• Ephemeral Credential: An ephemeral credential is a short-lived secret or token issued for a specific task or time window. It reduces the usefulness of theft, but it only improves security when issuance, renewal, and revocation are controlled tightly and tied to a verified workload identity.
• Identity Blast Radius: Identity blast radius is the amount of damage a compromised identity can cause before it is detected or revoked. For non-human identities, it depends on privilege scope, credential lifetime, and how many downstream systems trust the credential without additional checks.

Deepen your knowledge

Zero trust for non-human workloads and machine identity lifecycle governance are core topics in our NHI Foundation Level course, the industry's only accredited NHI security programme. If your programme is moving from static secrets toward runtime authorisation, it is worth exploring.

This post draws on content published by Aembit: Zero Trust for Non-Human Workloads. Read the original.