AI data security readiness is lagging enterprise adoption

At a glance

What this is: This report says AI adoption is already mainstream in enterprises, but visibility into how AI is being used remains very low.

Why it matters: IAM, IGA, PAM, NHI, and AI governance teams need to treat AI usage as an access and accountability problem, not just a data or model problem.

By the numbers:

• 83% of enterprises already use AI in daily operations.
• only 13% have strong visibility into how it’s being used.

Context

AI data security readiness is the gap between how quickly enterprises adopt AI and how slowly governance catches up. In this report, Cyera shows that usage is already widespread, but visibility, accountability, and control are still lagging.

For identity teams, the issue is not just model safety or data classification. It is whether AI usage can be governed with the same rigour applied to users, service accounts, and sensitive data access paths.

Key questions

Q: How should teams govern AI tool access without slowing adoption?

A: Use the same identity lifecycle controls you already apply to critical enterprise systems, then automate provisioning, role updates, and revocation so governed access is faster than ad hoc approvals. The goal is not to create a special AI process, but to make approved access the path of least resistance.

Q: Why do shadow AI tools create more identity risk than ordinary application sprawl?

A: Shadow AI often brings hidden identity relationships with it, such as OAuth grants, embedded secrets, and delegated API permissions. Those connections can create durable access to data and workflows even when the application itself looks harmless. The risk is not just the tool, but the ungoverned entitlement path behind it.

Q: What signals show that an AI governance programme is not working?

A: Warning signs include disconnected models built by different teams, repeated disputes over data ownership, inconsistent approvals and outputs that cannot be explained to stakeholders. If the organisation cannot trace which data supported a decision or who approved the model, governance is already failing at the operating level.

Q: How do IAM, PAM, and data security need to work together for AI?

A: They need to produce one control story for each AI use case: who approved it, which identities or tokens it uses, what data it can reach, and how that access is reviewed. When those answers live in separate systems with no shared evidence, AI governance becomes fragmented and weak.

Technical breakdown

AI adoption creates a visibility problem before it becomes a security problem

When AI enters daily operations faster than governance can track it, the first failure is usually visibility. Security teams lose the ability to answer basic questions: which systems are using AI, which data they can touch, and which identities or workflows authorize that access. That is why AI becomes a shadow identity issue as much as a data issue. Without reliable inventory, control design starts from guesswork, and every downstream policy, review, or detection rule inherits that blind spot.

Practical implication: establish an authoritative inventory of AI-enabled workflows, the identities they use, and the data they can reach.

Shadow AI behaves like an ungoverned access path

Shadow AI is not just unsanctioned software. In governance terms, it is an access path that can initiate requests, retrieve data, and influence outcomes without clear ownership or review. That makes it behave more like an unmanaged non-human identity than a conventional application. The risk is not limited to misuse by insiders. It also includes access drift, untracked integrations, and data exposure through systems that were never brought into the normal identity lifecycle.

Practical implication: apply identity ownership, approval, and review requirements to AI access paths the same way you would for other non-human identities.

Readiness gaps show up at the intersection of identity, access, and data control

AI governance fails when teams treat identity controls, access policy, and data protection as separate programmes. In practice, AI systems can combine broad data reach with weak attribution and inconsistent approval boundaries, which makes access decisions hard to explain after the fact. That is why AI readiness is a cross-domain problem, not a single control gap. The most material question is whether the organisation can prove who or what accessed which data, under what authority, and for what permitted purpose.

Practical implication: align IAM, PAM, and data governance teams around shared evidence for AI access, not separate control reports.

NHI Mgmt Group analysis

AI readiness is now an identity governance problem, not just a data governance problem. When enterprises adopt AI faster than they can inventory and attribute its access, the control failure starts at identity rather than at the model. The most useful framing is not whether AI is present, but whether every AI-enabled access path has a known owner, scope, and review point. Practitioners should treat AI access as part of the identity estate, not as an adjacent technology concern.

Shadow AI behaves like an unmanaged non-human identity. The report’s shadow AI framing matters because untracked AI usage creates the same governance pressure as unmanaged service accounts or stale tokens: no reliable owner, no durable approval record, and no dependable offboarding trigger. That is the practical difference between experimentation and governable deployment. Teams need to recognise that AI without clear attribution is operationally indistinguishable from other NHI blind spots.

Readiness metrics must move from adoption counts to governability evidence. Counting how many teams use AI says little about whether the organisation can control it. The stronger measure is whether security can show access provenance, data scope, and accountability for each AI use case. This shifts the programme from curiosity about adoption to evidence of control, which is what auditors, regulators, and incident responders will actually need.

AI governance will converge with NHI and privileged access governance. Once AI systems can reach sensitive data and trigger downstream actions, the relevant controls look less like generic AI policy and more like identity lifecycle, privilege scoping, and access review discipline. That convergence does not make AI identical to a service account, but it does mean the same governance mechanisms will increasingly be reused across both. Practitioners should plan for a single control model that spans human, machine, and AI access.

Only 13% visibility is a warning about control debt, not adoption maturity. A large majority of enterprises already using AI while a small minority can see how it is being used means organisations are accumulating governability debt faster than they can pay it down. The issue is not future readiness. It is whether current IAM, data security, and oversight processes can still explain who is acting, what they can reach, and why that access exists.

From our research library:

• Only 5.7% of organisations have full visibility into their service accounts, according to the Ultimate Guide to NHIs.
• Business leaders plan to spend $124 million on average on AI in 2026, and 91% say data security and risk will shape their AI strategy.
• Read next: Agentic AI Identity Risk Board Briefing

What this signals

AI adoption is outpacing governability. The article’s core warning is that enterprises are moving faster on deployment than on control evidence, which means AI readiness should be assessed through visibility and accountability, not enthusiasm. Security leaders should expect audit and incident questions to focus on who authorised the access path, not just which model was used.

Shadow AI is the practical shape of the problem. Once AI systems can access data without clear ownership, they behave like unmanaged non-human identities inside the environment. That pushes AI governance into the same operational territory as NHI lifecycle management, where inventory, attribution, and offboarding matter as much as policy language.

For practitioners

• Inventory AI access paths Catalogue every AI-enabled workflow, integration, and delegated access path, then map the identities, datasets, and business owners attached to each one.
• Assign accountable owners Require a named business and technical owner for each AI use case so approval, review, and incident response do not rely on informal team knowledge.
• Extend access reviews to AI usage Add AI-enabled systems to recertification and exception review cycles so changes in scope, data access, and use case purpose are visible.
• Align identity and data controls Join IAM, PAM, and data security evidence so teams can prove which identities, privileges, and datasets support each AI deployment.

Key takeaways

• AI adoption is already mainstream, but many enterprises still cannot see how AI is being used well enough to govern it confidently.
• The gap is operational, not theoretical, because hidden AI access behaves like an unmanaged identity path into sensitive data and workflows.
• Identity, access, and data teams need shared evidence for AI usage, or readiness will continue to lag deployment.

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.
• AI readiness: AI readiness is the state where an organisation can deploy AI systems without losing control of identity, access, and auditability. It goes beyond adoption or enthusiasm and asks whether the environment can govern AI tools and agents across the full stack, including data, devices, and lifecycle processes.
• Identity Governance: Identity governance is the set of controls that defines who approves access, who owns it, how it is reviewed, and when it is removed. In practice, it turns identity management from a deployment task into a durable control system that can withstand audits, organisational change, and operational growth.
• Access Attribution: Access attribution is the ability to link a data action to a specific human, service account, workload, or agent. It is essential for auditing, investigations, and policy enforcement because controls lose precision when the acting identity cannot be reliably identified.

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.