Cisco Duo and Britive extend zero trust to human and agentic access

At a glance

What this is: This is an analysis of how Duo and Britive connect authentication with runtime privilege control so human and agentic identities operate with zero standing privilege.

Why it matters: It matters because IAM teams now have to govern privilege duration, not just login assurance, across people, workloads and AI agents.

Context

Zero standing privilege is the idea that access should exist only for the duration of a task, then disappear. In identity programmes built around static roles and persistent credentials, that assumption breaks as soon as workloads and AI agents can act continuously across cloud and SaaS environments.

Britive's article treats the Duo integration as a shift from authentication-only Zero Trust to runtime privilege governance. That is the right lens for NHI and agentic AI security: the control point is no longer just who enters, but what access survives after entry and how quickly it is removed.

The article also ties this to auditability and least privilege across AWS, Azure, GCP, SaaS, Kubernetes and on-prem systems. That makes the topic broader than a single product integration because it sits at the junction of identity assurance, privilege containment and runtime enforcement.

Key questions

Q: What breaks when security teams keep standing privilege for AI agents and workloads?

A: Standing privilege breaks the assumption that access is temporary and reviewable. Once an agent can hold credentials across multiple tool calls, the blast radius expands, offboarding becomes unclear, and privilege persists beyond the task that justified it. The result is a governance gap, not just a technical weakness.

Q: When do ephemeral credentials reduce risk, and when do they not?

A: They reduce risk when issuance is tied to verified workload identity, strict scope, and short validity. They do not help much if the wrong principal can obtain them, if the scope is too broad, or if downstream services still trust the session after the original check. Short duration is useful, but only with strong governance.

Q: What are the signs that runtime privilege control is not working?

A: Warning signs include long-lived tokens, access grants that outlast the task, inconsistent revocation, and agents or workloads that keep permissions after their purpose is complete. If audit logs show persistent access rather than task-bound issuance, the runtime control is failing.

Q: How should teams govern non-human identities in AI-heavy environments?

A: Teams should govern non-human identities the same way they govern other privileged assets: assign ownership, minimise scope, rotate credentials regularly, and monitor for abnormal use. The key difference is speed. AI-driven workflows can exploit exposed access quickly, so detection and revocation must be automated and tied to lifecycle controls.

Technical breakdown

How runtime privilege differs from login authentication

Authentication proves an identity at the point of entry. Runtime privilege control governs what that identity can do after entry, for how long, and under which policy conditions. In this model, Duo handles user, device and contextual trust, while a separate control plane issues ephemeral credentials for the specific task. The architectural distinction matters because authentication alone cannot prevent privilege accumulation once a session, token or delegated workflow is already active. Zero standing privilege is therefore a runtime design choice, not an extension of MFA. It changes the control boundary from access initiation to access duration and revocation.

Practical implication: separate entry assurance from runtime authorisation and treat privilege duration as a first-class control objective.

Why ephemeral credentials reduce blast radius in NHI and agentic workflows

Ephemeral credentials are short-lived access tokens or secret equivalents issued just in time and revoked when a task completes. They reduce blast radius because the identity never keeps reusable access longer than necessary, which limits lateral movement, accidental overreach and post-task reuse. For NHIs and AI agents, this is especially important because the actor can operate repeatedly and at scale without the natural offboarding boundary that human employment provides. The control works only when credential issuance, scope and revocation are tied to task completion and policy context, not to static role membership.

Practical implication: map the shortest workable access window to each task and revoke the credential automatically at task completion.

What shared signals and policy enforcement add to zero standing privilege

Policy enforcement becomes more useful when it can consume identity posture, device trust and environmental signals in real time. Shared signals from identity intelligence, session telemetry and policy engines let the runtime layer decide whether an access grant is still justified after entry. That matters for agentic AI because the system may call tools, query data and chain actions without pausing for human review. The result is not just finer-grained control but a traceable record of every grant, denial and revocation, which is essential for audit and incident response.

Practical implication: feed runtime authorisation with live identity and session signals so access decisions can change as context changes.

Threat narrative

Attacker objective: The objective is to turn legitimate access into persistent, reusable privilege that can be abused beyond the original task.

1. Entry occurs through legitimate authentication, where the user, workload or agent is trusted to begin a session.
2. Privilege escalation is avoided only if the runtime layer issues task-scoped access instead of preserving standing credentials.
3. Impact is contained when revoked credentials cannot be reused for later tool calls, downstream APIs or cross-environment access.

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.
• Salt Typhoon telecom intrusions 2025: Salt Typhoon breached US telecoms mainly with stolen logins, then harvested SNMP strings and TACACS/RADIUS keys to spread and persist for years.

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 standing privilege is the new control boundary for both human and machine identities. Once agents and workloads can operate continuously, the security question is no longer whether an identity can authenticate, but whether it should retain access after the task begins. That shifts the centre of gravity from login assurance to privilege duration, revocation and auditability. Practitioners should treat runtime access as the primary governance object, not an afterthought.

The assumption that least privilege can be fixed at provisioning time is collapsing under agentic behaviour. That assumption was designed for actors whose intent and scope are stable enough to model up front. It fails when an agent can call tools, query data and chain actions at runtime because the needed access is not fully knowable in advance. The implication is that privilege governance must move from static assignment to task-scoped authorisation.

Identity programmes that stop at authentication leave the hardest problem untouched. Duo-style entry controls reduce impostor risk, but they do not govern what happens once a legitimate session is active. Britive's framing is useful because it exposes the real gap: persistent privilege inside otherwise well-authenticated workflows. Security teams should read that as a warning that access assurance and privilege containment are separate control layers.

Auditable runtime enforcement is becoming a minimum requirement for AI governance, not a nice-to-have control. When AI agents can act on behalf of business users, the organisation needs a record of what was accessed, for how long, and under whose accountability. That aligns with NIST AI Risk Management Framework style governance for AI oversight, but the operational primitive is still identity and privilege telemetry. The practical conclusion is that observability must extend to credential issuance and revocation, not just alerts.

Zero standing privilege for agentic AI should be treated as a governance pattern, not a single product feature. The article points to a broader market direction in which identity, policy, observability and workload access control converge. That convergence will force teams to re-evaluate where ownership sits between IAM, PAM, cloud security and AI governance. Practitioners should plan for cross-domain policy, not isolated point controls.

From our research library:

• 90% of IT leaders say properly managing NHIs is essential for a successful zero-trust implementation, according to the Ultimate Guide to NHIs.
• Only 5.7% of organisations have full visibility into their service accounts, according to the Ultimate Guide to NHIs.
• Read next: AI Agent Identity Security Buyer's Guide

What this signals

Identity assurance is no longer enough on its own: programmes that stop at authentication leave persistent privilege untouched, which is where most runtime abuse now concentrates. The control objective should shift toward task-bound access, revocation and traceability across humans, workloads and agents.

Zero standing privilege is becoming the practical Zero Trust test for AI agents: if an agent can keep credentials after the task, the architecture still has standing exposure. That is why access issuance and revocation must be observable as part of the security model, not hidden inside the workflow.

Runtime governance must be designed for identities that act faster than review cycles: access reviews assume privilege lasts long enough to be inspected, but agentic workflows can create and discard access within one task. The implication is that governance moves to the moment of issuance, not the next certification cycle.

For practitioners

• Enforce task-scoped access for every non-human identity Replace standing secrets and persistent role grants with access that is issued per task and revoked automatically when the task ends. Apply this to service accounts, workload identities and AI agents that can call tools or APIs.
• Separate entry authentication from runtime privilege decisions Keep MFA, device trust and SSO as entry controls, but add a distinct runtime authorisation layer that decides what the identity can do after login and how long the grant lasts.
• Log every credential checkout and revocation Capture grants, denials, session context and revocations in a system that SOC and IAM teams can review together. This creates an audit trail for human and agentic access paths.
• Map agent ownership to accountable humans Require every agent or automated workflow to have a named business or technical owner so runtime decisions can be traced back to a responsible party when access is abused or mis-scoped.

Key takeaways

• The article frames zero standing privilege as the missing layer between login assurance and runtime control for human, workload and agentic identities.
• The key operational change is that credentials should exist only for the duration of the task, then be revoked automatically.
• For practitioners, the main lesson is to govern privilege duration and observability alongside authentication, or persistent access will remain the dominant exposure.

Key terms

• Zero Standing Privilege: A control model in which an identity does not keep persistent access unless it is actively needed. For NHIs, this means credentials and permissions are issued for a narrow task and then removed. It reduces the time window and reuse value of stolen access.
• Ephemeral Credentials: Ephemeral credentials are short-lived access artefacts issued for a limited task or session. They reduce the window for abuse, but they only improve security when paired with strong scope limits, telemetry, and automatic revocation at task completion.
• 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.
• Identity Assurance: The confidence an organisation has that a person or system is truly who it claims to be before access or action is granted. In modern IAM, assurance depends on evidence quality, channel trust, and the strength of verification around high-risk decisions.

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.