TL;DR: C1.ai explains that modern environments now include humans, service accounts, and AI agents, with each agent and tool call adding credentials and permissions that quickly multiply governance complexity; ten agents calling one hundred tools can create 10,000 credentials to manage. The governance assumption that identities are mostly human and reviewable on a stable cadence no longer holds when machine and agent identity footprints grow faster than traditional IAM processes.
At a glance
What this is: This is a governance explainer showing that non-human identities now include service accounts and AI agents, and that each layer adds credentials, permissions, and review burden.
Why it matters: It matters because IAM teams now have to govern humans, workloads, and agents together, or risk losing visibility into over-permissioned access and stale credentials.
👉 Read C1.ai's explanation of how human and non-human identities are changing IAM governance
Context
Modern identity governance no longer stops at employees and contractors. In cloud-heavy environments, the real problem is that service accounts, workload credentials, and AI agents all participate in access decisions and execution, which makes identity sprawl a governance issue rather than just an inventory issue.
C1.ai's article argues that the traditional human-centric IAM model is incomplete because non-human identities now access resources, trigger workflows, and call tools at machine speed. That changes how practitioners think about authentication, authorisation, monitoring, and periodic review across the full identity estate.
Key questions
Q: What breaks when non-human identities are governed like human users?
A: Lifecycle triggers, ownership, and review processes stop working because machine identities do not generate joiner, mover, or leaver events. Access can persist after the original purpose disappears, leaving valid credentials outside normal certification paths. That creates a blind spot where privileged access remains active even though nobody can clearly explain why it still exists.
Q: Why do AI agents make non-human identity governance harder?
A: AI agents make governance harder because they can request tools, act autonomously, and change behaviour across sessions while still relying on machine credentials. That increases the number of access paths security teams must supervise. The result is a stronger need for task-scoped access, explicit ownership, and continuous monitoring of what the agent can reach.
Q: How do security teams keep non-human access from multiplying without control?
A: Start by inventorying every non-human identity, then tie each one to a named owner, a defined purpose, and a review date. Next, track which tools and systems each identity can reach so permissions can be reduced where access is broader than the task actually requires.
Q: Why do service accounts and AI agents need different controls from human users?
A: Service accounts and AI agents authenticate and act without the predictable patterns that human identity systems expect. They can operate across runtimes, scale quickly, and carry permissions into automated workflows. That means access decisions should consider workload context, runtime behaviour, and time-bound authority rather than relying only on user-centric IAM patterns.
Technical breakdown
How AI agents change the identity model
AI agents are not just automated scripts. In this article's framing, they ingest data, decide when to trigger actions, and call other tools, which means they behave like identities with runtime agency rather than passive workloads. Each agent may hold credentials, and each tool it calls can introduce a separate identity boundary. That creates a chain of authentication and authorisation events that traditional human IAM was never designed to count at this velocity or volume.
Practical implication: model agents as governed identities with explicit ownership, scope, and review, not as generic automation.
Why tool-to-tool chains expand credential risk
The article's tool-chain example shows how one agent can fan out into many downstream systems. Each additional tool call may require its own credential, token, or service account, so the number of trust edges grows faster than the number of visible users. The real technical issue is not just volume but coupling: credentials, permissions, and workflow logic become distributed across multiple systems, making it harder to know which access is still required.
Practical implication: map every agent-to-tool dependency so you can see where credentials are inherited, duplicated, or left standing after use.
What governance looks like across humans and non-humans
The article groups humans, service accounts, AI agents, tools, and systems into one governance domain because all of them must be authenticated, authorised, monitored, and periodically reviewed. That is the key architectural shift. Identity governance becomes a lifecycle problem across actor types, with different subjects but the same control intent. The challenge is to keep the governance model coherent when some identities act at human pace and others operate continuously or dynamically.
Practical implication: build one governance model that spans joiner, mover, leaver, and review processes across human and non-human identities.
Threat narrative
Attacker objective: The objective is to exploit excessive or unmanaged non-human access to reach systems, data, or workflows that should not be reachable under least privilege.
- Entry occurs when an over-permissioned service account, agent credential, or stored secret is present in the environment and can be reused beyond its intended scope.
- Escalation follows when AI agents or tool chains inherit broader permissions than the task requires, increasing the chance of unintended access paths.
- Impact comes from stale, forgotten, or insecurely stored credentials enabling misuse, compromise, or silent expansion of access across systems.
Breaches seen in the wild
- Scania insurance portal breach 2025: An attacker used an external user login, likely stolen by infostealer malware, to take insurance claim documents from a Scania portal.
Read and download The State of NHI & AI Agent Breach Report 2026, covering 150+ breaches impacting Non-Human Identities including AI Agents.
NHI Mgmt Group analysis
Identity governance has crossed a threshold where human-first models are no longer sufficient. The article's central claim is that identity estates now include service accounts, tools, and AI agents in the same control plane as employees. That means governance has to shift from a people-only programme to an identity programme that spans every actor type that can authenticate and act. Practitioners should treat this as a redesign of scope, not a minor control extension.
AI agents create credential multiplication, not just more automation. The ten-agents-by-one-hundred-tools example is not a curiosity, it is a signal that agentic systems can dramatically expand the number of credentials, trust boundaries, and review objects. The named concept here is identity footprint multiplication: every added agent and tool call increases the number of access relationships that must be owned and governed. Security teams should recognise that the review burden scales non-linearly.
Service accounts and AI agents belong to the same governance conversation because both can outlive the context that justified their creation. A service account may become stale, while an AI agent may continue to call tools long after its original workflow changed. The practical implication is that lifecycle governance, ownership, and periodic review must cover both classes with equal seriousness. IAM and IGA teams should stop treating non-human access as an exception path.
The real governance gap is not authentication alone, but the lack of end-to-end visibility across delegated access chains. When an agent calls a tool that calls another system, accountability fragments across multiple identities and systems. That makes monitoring and attestation much harder than in human IAM, where the subject is usually stable and easier to classify. Practitioners should assume delegated chains will be the default architecture, then govern them as such.
Least privilege remains the right principle, but its unit of control is now broader than a single account. The article shows that non-human identity risk emerges from the relationship between identities, tools, and workflows. That is why governance teams need to look at entitlement scope, stored credentials, and review cadence together rather than separately. The programme implication is straightforward: identity governance must become relationship-aware, not account-aware only.
From our research library:
- NHIs outnumber human identities by 25x to 50x in modern enterprises, according to the Ultimate Guide to NHIs.
- Read next: Ultimate Guide to NHIs
What this signals
Identity footprint multiplication: agentic systems do not just add another identity class, they multiply the number of credentials and trust edges that IAM teams must govern. The practical effect is that identity programmes will be judged by how well they can enumerate and review machine-to-machine access, not just human accounts.
As agentic workflows spread, the centre of gravity moves from login events to delegated access chains. That means practitioners should expect ownership gaps, stale permissions, and hidden tool dependencies to become the dominant failure modes unless governance expands across the full identity stack.
For practitioners
- Inventory all non-human identities Build a single register for service accounts, AI agents, tool identities, and any supporting credentials so ownership is not fragmented across teams.
- Map agent-to-tool credential chains Document which agent uses which tool, which credential each call depends on, and where permissions are duplicated across systems.
- Extend access reviews beyond employees Include service accounts and AI agents in periodic review cycles, with explicit approval for continued access and clear business ownership.
- Tighten credential hygiene in pipelines Remove embedded secrets from code and build pipelines, then shift to controlled issuance and monitored use for machine access.
- Apply least privilege to agent workflows Constrain each agent and supporting tool identity to the smallest workable scope, then revoke permissions that are only needed during setup or testing.
Key takeaways
- The article's core warning is that identity governance now has to span humans, service accounts, tools, and AI agents together, not as separate programmes.
- Agent-to-tool chaining can multiply credentials and permissions far faster than traditional IAM review cycles are built to handle.
- The most effective control response is a shared lifecycle model with ownership, least privilege, and periodic review for every identity that can act.
Standards & Framework Alignment
This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.
OWASP Non-Human Identity Top 10, OWASP Agentic AI Top 10 and MITRE ATT&CK address the attack and risk surface, while NIST CSF 2.0 sets the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| OWASP Non-Human Identity Top 10 | NHI-05 — Overprivileged NHI | The article centres on broad permissions and expanding access scopes for non-human identities. |
| NHI-07 — Long-Lived Secrets | The post highlights credentials that persist across service accounts, agents, and pipelines. | |
| Recommendation — Review NHI permissions for overbreadth and reduce any access that exceeds the identity's current task. Track and shorten credential lifetimes wherever machine identities keep access longer than required. | ||
| OWASP Agentic AI Top 10 | ASI03 — Identity & Privilege Abuse | AI agents in the article are granted tool access and decision paths that can exceed intended scope. |
| Recommendation — Constrain agent identity and privilege paths so runtime access cannot exceed approved task scope. | ||
| NIST CSF 2.0 | PR.AA-05 — Access Permissions, Entitlements and Authorizations | The article is about how permissions and authorisations must span human and non-human identities. |
| Recommendation — Apply entitlement governance to every identity type and remove access that lacks an active business need. | ||
| MITRE ATT&CK | TA0006;TA0008 — Credential Access; Lateral Movement | The article discusses credentials, tool chains, and permission spread that can enable abuse and movement. |
| Recommendation — Map non-human credential sprawl to credential-access and lateral-movement risks in your detection coverage. | ||
Key terms
- Non-Human Identity (NHI): A digital identity assigned to a non-human entity such as a software application, service account, API key, bot, machine, or AI agent that enables it to authenticate and interact with systems without direct human involvement. NHIs now outnumber human identities in most enterprises by 25 to 50 times.
- Agentic Identity: An agentic identity is a non-human identity used by an autonomous system that can act, call tools, and access data with execution authority. It needs the same governance discipline as other privileged identities, plus runtime context, ownership mapping, and revocation paths.
- Credential Sprawl: Credential sprawl is the uncontrolled accumulation of machine secrets, keys, and tokens across systems, teams, and environments. It usually starts with a single use case and ends with overlapping permissions, unclear ownership, and a larger attack surface than the organisation expected.
- Identity footprint: An identity footprint is the full set of accounts, roles, tokens, SSO links, and integrations created by a system or application. For SaaS, it shows where access lives after procurement ends and helps teams prove that retirement actually removed reachable access.
What's in the full article
C1.ai's full blog post covers the practical detail this post intentionally leaves for the source:
- How C1.ai distinguishes human identities, service accounts, tools, and AI agents in day-to-day governance
- Examples of the access patterns that create credential multiplication across agentic workflows
- The article's plain-language breakdown of why non-human identities need different review and monitoring habits
- The broader framing C1.ai uses to explain why the identity stack is now changing shape
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.
Published by the NHIMG editorial team on June 7, 2026.
Updated on October 7, 2026.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org