SailPoint-Entro acquisition shifts NHI governance for agentic AI

At a glance

What this is: SailPoint’s acquisition of Entro Security brings NHI discovery, credentials security, and agentic identity governance into one enterprise identity platform.

Why it matters: It matters because IAM teams now have to evaluate whether their human, machine, and AI agent governance remains fragmented or can support unified lifecycle control across all three.

By the numbers:

• 80% of organisations report their AI agents have already performed actions beyond their intended scope, including accessing unauthorised systems (39%), inappropriately sharing sensitive data (31%), and revealing access credentials (23%).
• 92% agree governing AI agents is critical to enterprise security, yet only 44% have implemented any policies to do so.

👉 Read SailPoint's acquisition announcement covering Entro Security and agentic identity governance

Context

SailPoint’s acquisition of Entro Security is a signal that agentic AI identity, NHI discovery, and credentials governance are now converging into one operational problem. The core issue is not whether organisations can see more identities, but whether they can govern human, machine, and agent access with the same lifecycle discipline.

For identity teams, the pressure point is the gap between broad entitlement governance and the granular reality of secrets, tokens, certificates, and AI tool use. The article frames that gap as a control-plane problem, where discovery without lineage and lifecycle control does not stop credential abuse or posture drift.

This is most relevant for programmes that already manage service accounts, workload identity, and access certification separately, because the combined scope now spans all three. That makes the post useful beyond one vendor transaction: it describes where identity governance is heading as machine and agent populations expand.

Key questions

Q: How should security teams govern AI agents that use secrets and tools at runtime?

A: Security teams should treat AI agents as governed identities with explicit ownership, approved tool scope, and continuous monitoring of behaviour while they are active. Static access approval is not enough when the agent can choose actions at runtime. The control model should combine lifecycle review, secret lineage, and runtime detection so that use of credentials stays attributable and bounded.

Q: Why do NHI programmes need secret lineage, not just secret discovery?

A: Secret discovery tells you what exists, but secret lineage tells you what is actually using the credential and for what purpose. That matters because hidden use inside codebases, pipelines, and containers can keep risky access alive even when the inventory looks clean. Governance only becomes actionable when each secret can be tied to a live workload, application, or agent.

Q: What breaks when lifecycle reviews are the only control for non-human identities?

A: Lifecycle reviews miss active misuse because they operate on periodic cycles, while credential abuse, malicious tool calls, and posture drift can happen during execution. In practice, that means an identity can be approved and still behave unsafely before the next review. Teams need runtime visibility alongside certification and offboarding to close that gap.

Q: Who should own governance when humans, machines, and AI agents share one identity fabric?

A: Ownership should sit with the programme that can connect accountability, entitlement scope, and runtime behaviour across all actor types. Separate teams can still manage their specialties, but the governance model needs one control plane for lineage, approval, and monitoring. Without that, handoffs between human IAM, NHI, and agentic AI create gaps that attackers or misbehaving systems can exploit.

How it works in practice

Why NHI discovery and lineage need to be linked

NHI discovery tells you what non-human identities exist, while lineage tells you which application, script, workload, or agent is actually using a secret or credential. Those are different control problems. Discovery without lineage leaves teams with inventory but not accountability, and lineage without discovery leaves hidden identities outside governance scope. In developer environments, secrets often live in codebases, CI/CD systems, and container registries, so the practical challenge is mapping credential use back to an owner and a purpose. That mapping is what turns raw secret visibility into governance action.

Practical implication: build identity inventory and secret lineage together, not as separate programmes.

How runtime NHI defence differs from workflow governance

Workflow governance handles certifications, segregation of duties, and lifecycle reviews. Runtime defence handles token behaviour, anomaly detection, malicious tool calls, and prompt-related abuse while the identity is active. The article distinguishes these layers clearly: one is compliance-oriented and periodic, the other is behavioural and continuous. That distinction matters because a workload or agent can be fully certified and still behave unsafely during execution. For IAM and security teams, the architectural question is where control ends at approval and where active monitoring begins.

Practical implication: separate approval workflows from behavioural monitoring, and design controls for both.

Why autonomous AI agents change the identity control plane

Autonomous AI agents can choose actions and tool calls at runtime, which means identity is no longer just about static entitlement assignment. The control problem becomes one of governing delegated action paths that may change during execution, especially when an agent can chain tool use across systems. In that model, access is not simply consumed, it is orchestrated. That is why the article ties agentic governance to a broader identity fabric rather than to point controls alone. The technical shift is from managing credentials to managing decision-bearing identities.

Practical implication: treat AI agents as governed identities with runtime constraints, not as ordinary automation.

NHI Mgmt Group analysis

Unified identity governance is now a market requirement, not a roadmap ambition. The acquisition reflects a broader shift in which human, machine, and agent identities can no longer be managed as separate programmes. Organisations that keep discovery, lifecycle, and runtime defence in different silos will miss the control relationships between secrets, workloads, and agent actions. The practitioner conclusion is straightforward: identity architecture is moving toward a single governance fabric across all actor types.

Deep secret lineage is becoming as important as broad identity visibility. Knowing that an NHI exists is no longer enough when the real risk sits in where that secret is embedded and what downstream resources it can activate. The specific value here is the ability to trace a credential back to an application, script, or agent in use, which closes the gap between inventory and accountability. The practitioner implication is to prioritise lineage over raw count metrics when judging governance maturity.

Runtime identity abuse is the control gap that lifecycle processes alone cannot close. Certifications, SoD checks, and access reviews still matter, but they operate on periodic assumptions that do not catch malicious tool calls or anomalous token behaviour in flight. Ephemeral credential trust debt: this is the growing gap created when organisations keep issuing short-lived secrets without proving how they are used during active execution. The practitioner conclusion is that runtime governance must complement, not replace, lifecycle control.

Agentic AI exposes where identity programmes still assume the actor is passive. The article’s framing shows that autonomous identities can create and combine tool calls at runtime, which means the governance model must account for behaviour that is not fully knowable at provisioning time. That is where AI agent governance and NHI governance converge: both must control delegated access, but autonomous actors add decision-making variance. The practitioner conclusion is to test whether current identity assumptions still hold once the actor can choose its own next action.

The market is heading toward governance platforms that combine breadth with depth. Broad identity graphs without technical NHI controls are incomplete, and technical secret scanners without lifecycle context are also incomplete. The direction of travel is toward platforms that can connect human accountability, machine lineage, and agent activity under one control model. The practitioner conclusion is to re-evaluate tool sprawl, because fragmented controls will increasingly look like a governance gap rather than a design choice.

From our research:

• 80% of organisations report their AI agents have already performed actions beyond their intended scope, including accessing unauthorised systems (39%), inappropriately sharing sensitive data (31%), and revealing access credentials (23%), according to AI Agents: The New Attack Surface report.
• Only 52% of companies can track and audit the data their AI agents access, leaving 48% with a compliance and breach investigation blind spot according to AI Agents: The New Attack Surface report.
• That governance gap becomes more visible when teams compare agent behaviour with the NHI lifecycle discipline in the NHI Lifecycle Management Guide, where ownership, offboarding, and review are tied to concrete identity state.

What this signals

Agentic identity governance will increasingly be judged by whether teams can connect approval, lineage, and runtime behaviour in one view. The new operating model is less about adding another discovery feed and more about proving that each secret, token, or agent action maps back to a current owner and a current purpose. The NHI Lifecycle Management Guide remains the best reference point for lifecycle discipline, but agentic programmes now need to extend that thinking into execution-time monitoring.

Only 44% of organisations have any AI agent governance policies in place, so most programmes are still early in the control curve. That means identity teams should expect a phase of consolidation where NHI discovery, access certification, and agent monitoring become operationally linked rather than separately reported. The priority is to design controls that survive scale, not to wait for perfect inventory before acting.

Ephemeral credential trust debt: as agents, workloads, and developer secrets proliferate, the hidden risk is not just exposure but the growing debt created when access is issued faster than it is attributed and retired. Teams that align secret lineage with offboarding discipline will be better positioned to use the Top 10 NHI Issues as a practical planning tool for the next governance cycle.

For practitioners

• Map secrets to active downstream use Inventory where secrets, tokens, and certificates are embedded in codebases, CI/CD pipelines, and container registries, then tie each item to the application, script, or agent currently using it.
• Separate lifecycle control from runtime monitoring Keep certifications, segregation of duties, and access reviews as periodic controls, but add continuous monitoring for token behaviour, anomalous tool calls, and prompt-related abuse.
• Review AI agent governance as an identity problem Treat autonomous agents as governed identities with delegated actions, approved tools, and clear ownership, rather than as generic automation that can be managed only through workflow policy.
• Consolidate human, machine, and agent oversight Assess whether separate programmes for human IAM, NHI discovery, and AI agent governance are creating blind spots between ownership, lineage, and active use.

Key takeaways

• The acquisition signals a shift from fragmented NHI tooling toward unified governance across human, machine, and AI agent identities.
• Entro’s research theme is clear: AI agents, secrets, and runtime behaviour are already creating measurable control gaps that lifecycle reviews alone cannot close.
• Identity teams should reassess whether they can trace secret use, prove accountability, and monitor runtime activity within one operating model.

Key terms

• Non-Human Identity: A non-human identity is any digital identity used by software rather than a person. That includes service accounts, API keys, tokens, certificates, workloads, bots, and AI agents. Governance focuses on ownership, scope, lifecycle, and the secrets that allow the identity to act.
• Secret Lineage: Secret lineage is the ability to trace a credential from storage to active use. It links a token, key, or certificate to the workload, application, or agent that depends on it, which turns inventory into accountability and makes hidden access patterns visible.
• Runtime Identity Governance: Runtime identity governance is the set of controls that watch identity behaviour while access is in use. It goes beyond approvals and certification by monitoring tokens, tool calls, and anomalous actions, especially where machine and agent identities can change behaviour during execution.
• Agentic Identity: An agentic identity is an AI identity that can choose actions and tools at runtime rather than simply following a fixed script. The governance challenge is to bound delegated authority, preserve accountability, and detect behaviour that diverges from the intended task scope.

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.

This post draws on content published by SailPoint: SailPoint completes acquisition of Entro Security to secure agentic identities. Read the original.