Adaptive Trust exposes the limits of conditional identity governance

At a glance

What this is: This is an analysis of Adaptive Trust, a continuous identity governance model that replaces periodic access review with ongoing assessment and remediation across human, non-human, and AI agent identities.

Why it matters: It matters because IAM teams cannot separate identity governance from cybersecurity posture when non-human identities now dominate the attack surface and traditional review-based controls lag behind real risk.

By the numbers:

• Eighty-four percent of Fortune 500 companies score a D or worse for their cybersecurity efforts, according to Business Digital Index research analyzing companies across seven key security dimensions.
• With up to 71% of breaches now attributed to compromised credentials and identity-based attacks, the state of cybersecurity posture and the state of identity security are impossible to separate.
• Only 5.7% of organisations have full visibility into their service accounts.

👉 Read Clarity Security's analysis of Adaptive Trust and conditional trust limits

Context

Adaptive Trust is a continuous identity governance model that evaluates access in real time instead of waiting for periodic review cycles. The core problem it addresses is straightforward: identity security no longer operates in a world where access can be certified long after it was granted and still be considered current.

The article argues that Conditional Trust, built mainly for human identities and formal employment relationships, does not scale to service accounts, API keys, OAuth tokens, bots, and AI agents. That gap matters for IAM, NHI governance, and lifecycle management because the programme boundary no longer matches the actual identity estate.

In practical terms, this is a governance maturity story, not a tool feature story. Organisations that still depend on campaign-based reviews are trying to control an environment that moves faster than their review cadence can observe.

Key questions

Q: How should security teams move from periodic access reviews to continuous identity governance?

A: Start by treating reviews as assurance, not detection. Then connect identity telemetry, entitlement data, and remediation workflows so low-risk changes can be acted on automatically and higher-risk cases are triaged with context. The goal is to shorten the time between finding and fixing, especially for non-human identities that review cycles often miss.

Q: Why do non-human identities break conditional trust models?

A: Non-human identities break conditional trust because they outnumber human accounts, rarely appear in HR workflows, and often keep access long after the original purpose ends. A review model built for stable employment relationships cannot see those changes quickly enough, so risk accumulates between review cycles.

Q: What do organisations get wrong about just-in-time access for NHIs?

A: They often apply just-in-time thinking only to privileged human sessions and leave service accounts, API keys, and bots with standing access. That leaves the main attack surface intact. JIT only changes the security outcome when it covers the full identity estate and is paired with lifecycle offboarding.

Q: How do you know if continuous remediation is actually working?

A: Look for reduced dwell time between risk detection and entitlement change, fewer identities outside lifecycle ownership, and fewer stale permissions surviving the review cycle. If risks remain open until the next campaign, the programme is still operating as a periodic review process rather than a continuous control.

Technical breakdown

Conditional trust and its review-cycle limits

Conditional Trust relies on identity being granted access, then periodically checked through access reviews, audit logs, and separation of duties controls. That model works when the identity subject is stable and the organisation can tolerate delayed correction. It breaks when the environment contains many more non-human identities than human users, because the review cycle becomes a snapshot of historical state rather than a control over current risk. The architecture is fundamentally retrospective: it finds drift after the fact and depends on manual follow-through to fix it.

Practical implication: treat periodic access reviews as a compliance control, not as your primary risk-reduction mechanism.

Continuous assessment and connected remediation

Adaptive Trust combines assessment and remediation in one workflow. Activity logs, entitlement data, and expected-behaviour signals are evaluated continuously, then routed directly to a remediation action rather than a ticket queue. The important technical shift is not just monitoring more often, but collapsing the delay between finding and fixing. That changes governance from a campaign process into an always-on control plane, where low-risk actions can be automated and higher-risk decisions are surfaced with context.

Practical implication: connect detection outputs to automated or semi-automated remediation paths instead of relying on manual case management.

Just-in-time provisioning for full identity scope

Just-in-time provisioning removes standing permissions by granting access only for a defined task and revoking it when the task is done. In Adaptive Trust, that pattern applies across human users, contractors, service accounts, API keys, bots, and AI agents. The technical significance is scope: the same entitlement logic must govern identities that do not exist in HR systems and may never appear in traditional joiner-mover-leaver workflows. Without that broader scope, JIT remains partial and the residual standing access continues to create attack surface.

Practical implication: extend JIT design and lifecycle controls to non-human identities, not just privileged human accounts.

Threat narrative

Attacker objective: The attacker aims to convert stale identity governance into durable access that remains usable long enough to support lateral movement, data exposure, or operational disruption.

1. Entry occurs through standing credentials or stale access that remains valid between governance cycles, especially for service accounts and other non-human identities.
2. Escalation happens when the identity keeps privileges it no longer needs, allowing activity that should have been removed during review.
3. Impact is broader lateral movement, unauthorized system access, or persistence that survives long enough to be exploited before the next certification cycle.

Breaches seen in the wild

• Moltbook AI agent keys breach — Moltbook breach exposed 1.5M AI agent keys.
• AI LLM hijack breach — attackers used stolen AWS access keys to hijack Anthropic LLM models on Bedrock.

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

NHI Mgmt Group analysis

Adaptive Trust is the clearest response to the failure of review-based identity governance. Campaign-driven governance assumes risk can be captured, assigned, and resolved on a human schedule. That assumption no longer holds when non-human identities, third-party credentials, and AI agents change state faster than the review cycle can observe. Practitioners should read this as a structural shift in how identity control must operate.

Conditional Trust was designed for a narrower identity estate than most enterprises now defend. It was built around human identities with formal employment relationships, not service accounts, API keys, OAuth tokens, and bots that dominate modern access paths. The result is a governance boundary that excludes the primary attack surface. Practitioners should treat scope mismatch as the core defect, not a minor coverage gap.

Access review processes assume access persists long enough to be reviewed, but that premise weakens as remediation becomes more automated. If assessment and remediation remain separate, risk accumulates between workflows and review output becomes stale before action begins. That is why continuous remediation is not just faster governance, it is a different control model. Practitioners should reframe certification as verification, not primary protection.

The named concept here is identity blast radius: the amount of damage a stale entitlement can create before the next governance cycle catches it. Adaptive Trust tries to shrink that blast radius by shortening the gap between detection and correction across every identity type. The practical conclusion is that programmes should measure how long risky access remains live, not just whether it was eventually reviewed.

Cross-actor governance is now mandatory because the same lifecycle rules apply to humans, NHIs, and autonomous systems, but the evidence of control failure appears first in non-human identity sprawl. Once third-party access, service accounts, and agent identities sit outside lifecycle management, the rest of IAM inherits the blind spot. Practitioners should use that blind spot as a signal that the governance model itself has outgrown its original assumptions.

From our research:

• 80% of identity breaches involved compromised non-human identities such as service accounts and API keys, according to the Ultimate Guide to NHIs.
• 71% of NHIs are not rotated within recommended time frames, increasing the risk of compromise over time.
• Guide to the Secret Sprawl Challenge shows why hidden credentials in code and CI/CD remain a persistent governance failure point.

What this signals

Identity blast radius: the real governance question is no longer whether access was reviewed, but how long risky access remained live before remediation. That shift pushes IAM teams toward continuous control design, because campaign timing is too slow to manage service accounts, third-party access, and agent identities at enterprise scale.

With 96% of organisations storing secrets outside secrets managers in vulnerable locations, per the Ultimate Guide to NHIs, the practical signal is clear: discovery and remediation still have to catch up with the way identities are actually provisioned. Teams should expect more scrutiny on lifecycle ownership and closer integration between identity governance and secret management.

The programme implication is that access reviews will increasingly be treated as evidence of control operation, not proof of control effectiveness. That matters for both human IAM and NHI governance, because organisations that cannot remove standing access quickly will keep rediscovering the same risk in every cycle.

For practitioners

• Inventory every identity class in scope Build a single inventory that includes human users, contractors, third parties, service accounts, API keys, bots, and AI agents. Map the relationships between them so review, remediation, and ownership do not stop at HR-linked accounts.
• Collapse review findings into remediation workflows Route risk findings into direct remediation actions instead of ticket queues. Start with low-risk, high-confidence changes, then expand automation only after you can prove the system’s decisions match competent human judgment.
• Define where standing access must disappear Identify entitlements that should exist only for a task, then convert them to just-in-time access or removal by default. Focus first on service accounts and third-party credentials that outlive the work they were created for.
• Reclassify access reviews as assurance control Use certification cycles to confirm that continuous controls are working, not to discover the first instance of risk. If a review is the first time you learn an entitlement is wrong, the programme is still retrospective.
• Measure identity blast radius in days, not campaigns Track how long risky access remains active after it is detected and how many identity types are still outside lifecycle management. Those two measures show whether Adaptive Trust is becoming operational or staying theoretical.

Key takeaways

• Adaptive Trust replaces periodic review with continuous assessment and remediation across human and non-human identities.
• The scale of the problem is already visible in the data, with identity compromise and NHI exposure driving most modern access risk.
• IAM teams should measure how fast risky access is removed, not only whether it was eventually reviewed.

Key terms

• Adaptive Trust: A continuous identity governance model that evaluates access in real time and remediates risk as it appears. It extends governance beyond periodic reviews by tying assessment, context, and response together across human and non-human identities.
• Conditional Trust: An identity model that grants access and then checks it on a recurring basis through reviews, logs, and policy controls. It improves on inherent trust, but it still assumes risk can be understood and corrected on a schedule rather than continuously.
• Just-in-Time Provisioning: A pattern that grants access only for the duration of a specific task and removes it once the task is complete. For non-human identities, it is most effective when paired with lifecycle ownership, so standing permissions do not reappear outside the intended workflow.
• Identity Blast Radius: The amount of harm a risky entitlement can cause before governance detects and removes it. In modern identity programmes, this is shaped less by whether access exists and more by how long it remains live after it should have been withdrawn.

What's in the full article

Clarity Security's full blog post covers the operational detail this post intentionally leaves for the source:

• The step-by-step Adaptive Trust workflow for moving from assessment to remediation.
• The three maturity stages of remediation automation and how human-in-the-loop decisioning changes over time.
• The specific limitations of campaign-based IGA tooling when it is extended to non-human identities.
• The article's full explanation of how just-in-time provisioning fits into the target operating model.

👉 Clarity Security's full post covers the remediation model, lifecycle scope, and governance gaps in more detail.

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 responsible for identity security strategy or NHI governance in your organisation, it is worth exploring.