C1.ai report shows agentic AI and NHI risk rising together

At a glance

What this is: C1.ai's report shows security teams accelerating agentic AI adoption while treating NHI risk, identity attacks, and security pressure as concurrent problems rather than separate tracks.

Why it matters: IAM and NHI practitioners need to see that agentic AI does not replace machine identity governance. It raises the bar for access control, visibility, and runtime oversight across human, NHI, and autonomous actors.

By the numbers:

• 89% of respondents say they plan to implement AI agents in their security departments within the next two years.
• 82% of organizations experienced at least one identity-driven cyberattack or data breach in the past year.
• 93% of security leaders said the risks associated with NHIs are urgent.
• 84% of security leaders said their identity and access management budgets are increasing.

👉 Read C1.ai's report on agentic AI adoption and rising NHI risk

Context

C1.ai's 2025 Future of Identity Security Report describes an identity landscape in which agentic AI adoption, NHI sprawl, and identity-driven attacks are advancing at the same time. The practical question for security programmes is no longer whether to govern machine access, but how to do it when autonomous behaviour and standing machine credentials collide.

The report is based on 494 security professionals across industries including financial services, healthcare, technology, and manufacturing. That mix matters because it shows these pressures are not confined to one sector or one maturity band; they are becoming a mainstream identity governance problem for teams that must manage human access, service accounts, secrets, and AI agents together.

Key questions

Q: How should security teams govern AI agents and NHIs differently?

A: Security teams should govern NHIs as predictable machine identities and AI agents as runtime actors that can alter behaviour after authentication. That means static entitlements, inventory, and rotation remain central for NHIs, while agents need behaviour monitoring, delegation tracing, and ownership controls that account for tool choice and execution timing.

Q: Why do over-provisioned NHIs create such a large security risk?

A: Because non-human identities often operate continuously, connect to sensitive systems, and are difficult to review manually. When they have more privilege than the task requires, compromise of one credential can become rapid lateral movement. The risk comes from persistence, reach, and weak ownership, not just from the credential itself.

Q: What signals show that identity controls are not keeping up with agentic AI?

A: Look for tool calls that are hard to attribute, access that expands across multiple services in one session, and approvals that do not explain the eventual action chain. Those patterns show that static entitlements are no longer capturing runtime reality. A mature programme should be able to reconstruct the execution path from identity evidence.

Q: When should organisations prioritise NHI governance before scaling agentic AI?

A: They should prioritise it before deployment scales beyond a handful of agents, especially in fragmented environments where tools and logs are siloed. Once autonomous systems begin acting on sensitive data or high-risk workflows, governance becomes harder to retrofit. Building controls early lets teams scale AI without multiplying hidden credentials, access drift, and accountability gaps.

Technical breakdown

Why agentic AI changes the identity control model

Agentic AI is not just another automation layer. An AI agent can choose actions, invoke tools, and execute at runtime, which means the identity problem shifts from static provisioning to controlling what the actor is allowed to do in the moment. That breaks assumptions embedded in conventional IAM, where access is usually defined before execution begins. For security teams, the core issue is not whether an agent is useful. It is whether existing authorisation, logging, and approval paths can keep pace with independent runtime behaviour.

Practical implication: move governance decisions to issuance, tool scope, and runtime authorisation rather than relying on post hoc access review.

Why NHI sprawl remains the underlying control problem

Non-human identities include service accounts, API keys, tokens, secrets, and certificates. They become risky when they are long-lived, over-provisioned, poorly inventoried, or difficult to tie back to an owner and a business purpose. The report's findings fit a familiar pattern: as machine access expands, visibility alone is not enough if the credentials themselves persist too long or reach too far. That is why NHI governance sits underneath the AI conversation rather than beside it.

Practical implication: inventory every NHI, bind it to ownership and purpose, and reduce standing access to the minimum necessary scope.

How identity-driven attacks expose governance gaps across the stack

An identity-driven breach rarely starts with identity policy failure alone. It usually exploits weak credential lifecycle control, excessive privilege, or poor separation between human, machine, and delegated access. The report's 82% figure shows this is no longer a niche issue. When identity becomes the path of least resistance, the programme problem is not just prevention. It is the inability to see, classify, and constrain every actor that can authenticate and act on behalf of the business.

Practical implication: treat identity as the control plane for detection and containment, not just for authentication.

Threat narrative

Attacker objective: The attacker aims to abuse identity trust, expand access through machine or delegated credentials, and reach data or systems that ordinary application controls should have protected.

1. Entry begins when high-value machine credentials or agent access is introduced into production without a sufficiently tight ownership and scope model.
2. Escalation occurs when over-provisioned NHIs or agent permissions allow broader system access than the task actually requires.
3. Impact follows when identity-driven attacks or data breaches use those standing privileges to move across environments and reach sensitive assets.

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

Agentic AI turns identity from a provisioning problem into a runtime governance problem. The article's numbers point to a market that is adopting autonomous behaviour faster than its control model can evolve. Conventional IAM assumes access can be defined up front and later certified, but agentic systems can choose tools and act within the same session. The implication is that identity governance now has to reason about execution, not just entitlement.

Ephemeral intent is becoming the new identity boundary. Service accounts and API keys were already difficult because they outlived the human workflow that created them. Agentic AI compounds that problem by adding actors whose purpose can change mid-task and whose access needs may not be knowable at provisioning time. That is a different governance problem from classic overprovisioning, and it forces practitioners to control what the actor may do at runtime, not only who created it.

Identity blast radius is now the most useful way to think about NHI and agentic risk. The report shows teams are worried about both scale and urgency, which makes breadth of exposure more important than isolated misconfigurations. One over-privileged service account is a local issue; thousands of loosely governed NHIs and agents create systemic blast radius. Practitioners should treat every identity with no natural human review point as a containment problem, not just an access problem.

Security budgets are rising, but governance maturity will still lag unless programmes unify human, machine, and autonomous controls. The report shows spending pressure, but spend alone does not resolve the root issue if teams buy point tools for separate identity silos. Access governance, ownership, lifecycle offboarding, and runtime authorisation need to be managed as one discipline across all actor types. The practical conclusion is that identity architecture now has to be actor-aware by design.

What this signals

Identity security is shifting from static control to runtime governance. As autonomous behaviour becomes more common, the old assumption that access can be defined once and reviewed later becomes less useful. Programmes need controls that can bound tool use, data access, and execution scope while the actor is active.

Identity blast radius is now a programme-level risk metric. Teams should ask how many service accounts, secrets, and agents can reach sensitive environments without a human review point. The larger that unreviewed surface becomes, the more likely a single credential or agent workflow becomes an enterprise incident.

For practitioners

• Map identity controls by actor type Separate human IAM, NHI governance, and autonomous agent oversight into distinct control paths so each actor type is managed to its own lifecycle and runtime risk.
• Inventory and classify all NHIs Build an authoritative inventory of service accounts, API keys, tokens, and certificates, then tie each item to an owner, business purpose, and rotation expectation.
• Restrict agent tool scope at runtime Constrain which tools and data sources an agent can use in production, and treat any expansion of scope as a governance decision rather than a convenience setting.
• Reduce standing privilege across machine access Replace broad, persistent access with narrower entitlements and shorter-lived credentials so an identity-driven compromise has less room to move.
• Measure governance by blast radius Track how many identities can reach sensitive systems without a human review point, and use that number to prioritise remediation work.

Key takeaways

• The report shows that agentic AI adoption and NHI risk are climbing together, not as separate issues but as one expanding identity governance problem.
• Identity-driven attacks remain common enough that security teams are increasing IAM budgets, yet the control gap is still about ownership, privilege scope, and lifecycle visibility.
• The strongest response is to govern runtime access, reduce standing privilege, and classify every non-human actor by its actual role in production.

Key terms

• Agentic AI: Autonomous AI systems capable of planning, deciding, and taking actions, including calling APIs, writing code, and orchestrating other agents, with minimal human oversight. Agentic AI introduces new NHI risks as agents must authenticate to external services.
• 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.
• Identity Blast Radius: The amount of damage a compromised identity can cause across systems, data, and infrastructure. In NHI environments, it is shaped by permissions, network reach, and administrative capability rather than by the credential alone. Reducing blast radius is a containment strategy that limits lateral movement and data exposure.
• Standing Privilege: Standing privilege is access that remains active even when no immediate task requires it. For NHI programmes, it is a common failure mode because long-lived credentials and persistent roles create unnecessary exposure. Reducing standing privilege usually means tighter expiry, on-demand access, and clearer review of who or what still needs access.

What's in the full report

C1.ai's full blog covers the operational detail this post intentionally leaves for the source:

• The full survey breakdown behind the 494 security professionals and how the responses vary by industry.
• Additional detail on the reported rise in identity-driven attacks and how leaders are translating that pressure into budget decisions.
• The article's own framing of why security teams are adopting agentic AI despite the risks.
• Context on the specific NHI concerns the report surfaces, including over-provisioning and long-lived credentials.

👉 The full C1.ai post includes the survey framing, priority shifts, and security leader commentary behind the findings.

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.