AI agent visibility is now the first control for NHI governance

At a glance

What this is: This blog argues that AI agent sprawl has turned discovery and unified visibility into the prerequisite for governing non-human identities.

Why it matters: It matters because IAM and IGA teams cannot enforce scope, ownership, or access review on AI agents they cannot inventory, which leaves machine access paths outside governance.

Context

AI agent visibility is the governance gap when non-human identities are proliferating faster than teams can inventory them. The article frames a simple problem: if the organisation cannot tell where AI agents exist, it cannot assign ownership, map access paths, or govern the resulting machine identities.

In SailPoint's framing, shadow agents are not a future-state issue but an active control blind spot across browsers, endpoints, and SaaS applications. That makes discovery a baseline IAM and NHI requirement rather than a reporting exercise, because the access model has to start with an accurate inventory.

Key questions

Q: What breaks when AI agents are not inventoried or classified?

A: When AI agents are not inventoried or classified, organisations lose the ability to assign risk, apply the right obligations, and prove control to auditors. The result is hidden access, inconsistent oversight, and weak accountability when the agent touches regulated data or triggers business actions.

Q: Why do shadow agents create a bigger risk than ordinary automation?

A: Shadow agents create more risk because their authority can expand quietly as teams adapt them to new tasks. Ordinary automation usually has a narrower, more fixed control path. Shadow agents can cross systems, accumulate permissions, and continue acting without a clear owner, which makes accountability and containment much harder.

Q: How should security teams handle AI agent visibility?

A: Security teams must conduct an exhaustive discovery process to identify all deployed AI agents, both sanctioned and unsanctioned, across the organization. This visibility is foundational for effective governance and enables timely interventions against potential threats.

Q: Should organisations treat AI agents as NHI or as application automation?

A: They should treat them as governed non-human identities with application behaviour. That framing captures the need for identity, scope, lifecycle, and audit controls while still recognising that agents operate through APIs, tools, and workflows. Pure automation language hides the governance gap.

Technical breakdown

Why agent discovery is the first control

Discovery is the act of finding identities that already exist but are not represented in governance records. For AI agents, that means browser-based automation, endpoint scripts, and SaaS-embedded agents can all hold access without a durable inventory entry. Without discovery, there is no authoritative source for entitlements, owners, or review scope. The technical problem is not just missing tooling records. It is a broken identity map, where access may exist in runtime systems but not in the governance layer that should govern it.

Practical implication: inventory AI agents before you attempt access certification or policy enforcement.

Unified visibility turns a list into an access map

A list of AI agents is only a starting point. The governance value comes when the inventory connects each agent to a human owner, the machine identities it uses, the entitlements it holds, and the data it can reach. That creates an identity graph, which is a relationship model for understanding how access flows across human and non-human actors. In NHI programmes, the graph matters because access risk is usually created by combinations, not isolated credentials. A connected view exposes where ownership, privilege, and data sensitivity collide.

Practical implication: model human-to-agent-to-data relationships, not just agent counts.

Shadow AI changes the operating model for NHI governance

Shadow AI is unmanaged AI behaviour that sits outside formal identity controls. In the article's context, that includes agents operating invisibly in channels that traditional onboarding processes do not cover. The governance challenge is that these identities can be both ephemeral and persistent: ephemeral in how they are deployed, persistent in how long their access can remain unnoticed. That combination means NHI governance must move from periodic clean-up to continuous discovery with ownership attribution and scope mapping.

Practical implication: treat unidentified agents as governance gaps, not merely unsanctioned tooling.

Breaches seen in the wild

• CoPhish OAuth phishing via Copilot Studio: Datadog showed Copilot Studio agents on a Microsoft domain can front OAuth consent phishing and forward stolen tokens; no victims reported.
• AI agent retail card theft campaign 2026: AI agents breached 27+ retailers for about $25 each, used cloud keys and a Secrets Manager dump, and stole 600,000+ payment cards.

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

NHI Mgmt Group analysis

Visibility debt is now the first NHI governance failure mode: The article points to a control environment where AI agents exist faster than governance records are updated. That is not a monitoring problem alone. It is a structural identity gap, because any downstream control that depends on an inventory will fail when the inventory is incomplete. Practitioners should treat discovery as the prerequisite control that determines whether every other NHI process can operate at all.

Full-spectrum visibility is really identity-path visibility: Counting agents does not solve governance if the organisation cannot trace which human owns them, which machine identities they use, and which data they can touch. The article's strongest point is that access risk emerges along a path, not inside a single credential. That makes relationship mapping more valuable than point-in-time auditing for NHI programmes.

Shadow agents collapse the assumption that non-human identities are centrally introduced: NHI governance was designed for identities that at least enter the environment through recognisable onboarding and inventory processes. That assumption fails when agents appear inside browsers, endpoints, and SaaS workflows outside normal provisioning flows. The implication is that governance must be built to discover identities after deployment, not only at creation time.

Identity graphs will become the control plane for agentic governance: The article implicitly describes a shift from static inventory to contextual control. Once agent ownership, privilege, and data access are linked, teams can prioritise the identities that create the largest exposure surface. That is the difference between knowing an agent exists and knowing whether it matters.

NHI visibility is the gating function for real-time threat response: Real-time response only works when the responding team can identify which non-human identity is involved and what it can reach. The practical lesson is not merely to improve discovery rates. It is to make visibility the operating assumption behind every other identity security decision.

From our research library:

• Only 5.7% of organisations have full visibility into their service accounts, according to the Ultimate Guide to NHIs.
• 70% of organisations grant AI systems more access than they would give a human employee performing the exact same job, according to the 2026 Infrastructure Identity Survey.
• Read next: Agentic AI Identity Guide

What this signals

Discovery is becoming the control boundary for agentic identity programmes: Once AI agents can appear in browsers, endpoints, and SaaS workflows without a clean onboarding trail, periodic reviews stop being enough. Teams need continuous discovery tied to ownership and entitlement context, or their governance model will always trail the environment.

Identity graphs are the practical answer to agent sprawl: The important unit is not the agent alone but the path from human owner to machine identity to data access. That relationship model is what turns shadow AI into something IAM and IGA teams can actually govern.

70% of organisations grant AI systems more access than they would give a human employee performing the exact same job, according to the 2026 Infrastructure Identity Survey: that gap shows why visibility is now a control issue, not a reporting issue, because over-scoped access is easier to miss when the actor is invisible.

For practitioners

• Build a continuous AI agent inventory Track every agent across browser sessions, endpoints, SaaS applications, and major AI platforms so the governance record reflects runtime reality.
• Map ownership and access paths Connect each agent to a human owner, the machine identities it uses, its entitlements, and the sensitive data it can reach.
• Treat unidentified agents as governance exceptions Flag agents that cannot be tied to an owner or source system and route them into review before they accumulate standing access.
• Use discovery results to reset IAM priorities Use the agent inventory to decide which identities need immediate recertification, tighter entitlements, or removal from unknown channels.

Key takeaways

• AI agent sprawl creates an inventory problem first and a policy problem second, because governance cannot operate on identities that remain undiscovered.
• The article's core evidence is architectural rather than numerical: agents now appear across browsers, endpoints, and SaaS applications outside traditional provisioning paths.
• The immediate implication is to make discovery, ownership mapping, and access-path reconstruction the starting point for agentic NHI governance.

Key terms

• 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.
• Identity Graph: An identity graph is a relationship map that connects identities, assets, data, and permissions so teams can see how access actually flows. In NHI programmes, it helps explain which agent is related to which owner, which system, and which policy boundary.
• Discovery: Discovery is the process of finding AI agents and other non human identities already operating in an environment. It relies on telemetry from browsers, directories, and devices to build an inventory of tools, workloads, and access paths that may otherwise remain invisible to central IT.
• Agentic security: The practice of governing software actors that can choose actions, tools, and timing in production workflows. It extends identity, authorization, logging, and lifecycle control to agents so their behaviour is tied to a verifiable principal and a revocable permission set.

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 building or maturing an IAM programme, it is worth exploring.