NASCIO priorities show why identity-first government is now essential

At a glance

What this is: This is SailPoint’s interpretation of NASCIO’s 2026 state CIO priorities, arguing that identity-first governance is the control plane for AI, cloud, modernization, and access risk.

Why it matters: It matters because IAM, NHI, and emerging agentic AI controls now sit inside the same programme, and state agencies cannot modernise safely if identity governance lags behind access expansion.

By the numbers:

• 97% of NHIs carry excessive privileges, increasing unauthorised access and broadening the attack surface.
• 96% of organisations store secrets outside of secrets managers in vulnerable locations including code, config files, and CI/CD tools.
• 92% of organisations expose NHIs to third parties, raising concerns about supply chain security.

👉 Read SailPoint’s analysis of NASCIO’s 2026 identity-first priorities

Context

State agencies are being asked to modernise faster while also tightening control over cloud access, AI usage, and legacy system exposure. In practice, that means identity governance is no longer a back-office discipline. It is the control layer that determines whether new digital services, analytics platforms, and AI tools expand capability or simply widen access risk.

The article argues that a single governance fabric can span human users, service accounts, and AI agents across hybrid environments. That framing is useful because the operational failure mode is the same across each population: access expands faster than oversight, and review, deprovisioning, and policy enforcement fall behind the pace of change.

Key questions

Q: How should state agencies govern AI tools that can reach sensitive data?

A: State agencies should inventory every AI-connected access path, assign an owner, and require a revocation path before the tool is allowed to touch internal data. Governance should cover sanctioned assistants, shadow AI, plugins, and browser extensions. If the organisation cannot explain what the tool can see, the tool should not be connected to protected data.

Q: Why do cloud modernization programmes increase identity risk so quickly?

A: Cloud modernisation increases identity risk because it multiplies the number of permissions, service accounts, integrations, and delegated access paths that must be governed at once. If lifecycle controls do not keep pace, access accumulates faster than review and deprovisioning can remove it, creating lingering exposure across both new and legacy systems.

Q: What do security teams get wrong about shadow AI?

A: They often treat shadow AI as an application-usage problem when it is also an identity problem. The real issue is unmanaged access paths that can inherit permissions, reach sensitive data, and bypass normal approval workflows. Discovery must therefore include the identities and credentials behind the tools, not just the tools themselves.

Q: Who should be accountable for orphaned access in consolidation projects?

A: The accountable owner should be the programme or system team that controls the migration, because orphaned access is a lifecycle failure created by delayed offboarding and incomplete entitlement cleanup. Access should not survive system retirement unless an explicit business owner signs off on the exception.

Technical breakdown

Identity governance for humans, machines, and AI agents

Modern state environments no longer separate neatly into user access and machine access. Human users, service accounts, and AI agents all consume shared data and APIs, but they create different governance problems. Human access tends to be reviewed on role and employment status. Machine access depends on lifecycle state, credential scope, and rotation discipline. AI agents add runtime variability because the same identity may interact with more data, more tools, and more systems over time. A unified identity control plane matters because each actor type can carry excess privilege into the same cloud or legacy environment.

Practical implication: Treat identity governance as a cross-actor programme and map each access path to the correct lifecycle, review, and privilege controls.

Shadow AI and unapproved access paths

Shadow AI is not just an application risk. It is an identity problem because unsanctioned tools may connect to internal data through browser extensions, plugins, or delegated credentials without the agency knowing what data they can reach. Once those tools are connected, the risk is no longer limited to user behaviour. The hidden identity can persist, inherit permissions, and create an untracked route to sensitive records. This is why discovery and entitlement visibility must extend beyond approved applications to every AI-connected access path.

Practical implication: Inventory AI-enabled access paths the same way you inventory privileged accounts and revoke anything that cannot be tied to an owner and purpose.

Lifecycle drift in cloud and legacy environments

The article’s strongest operational point is that modernisation fails when access lifecycle processes do not keep pace with platform sprawl. In hybrid estates, permissions often linger after role changes, contractor offboarding, or system consolidation. That produces orphaned access, inconsistent entitlements, and delayed deprovisioning. The same pattern appears in both mainframe and cloud estates because the governance issue is not the platform itself. It is the lack of a consistent lifecycle model that follows access from creation through change to revocation.

Practical implication: Standardise joiner-mover-leaver controls across cloud, legacy, and shared service accounts so no environment escapes the same governance cycle.

Threat narrative

Attacker objective: The objective is to reach sensitive state data through poorly governed identities and turn access sprawl into data exposure or automated misuse.

1. Entry occurs when employees use shadow AI tools or when cloud services are added without centralised identity controls, creating untracked access paths into sensitive data.
2. Escalation follows as permissions expand over time, former staff retain access, and AI or machine identities accumulate unnecessary privileges across shared systems.
3. Impact is the exposure of citizen data, compliance findings, and the possibility of a broad automated breach if cloud-native AI tools inherit excessive access.

Breaches seen in the wild

• Cisco DevHub NHI breach — IntelBroker exploited exposed Cisco credentials, API tokens and keys in DevHub.
• Codefinger AWS S3 ransomware attack — Codefinger used compromised AWS credentials to encrypt S3 buckets via SSE-C.

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

NHI Mgmt Group analysis

Identity-first government has become a governance necessity, not a modernization preference. The article is right that cloud expansion, AI adoption, and legacy integration are converging on the same control problem: who and what can access data, when, and under whose authority. That is a classic NHI and IAM convergence issue, not a standalone platform issue. State agencies that treat identity as an enabling service will keep discovering that access decisions are the part of modernization that fails last and hurts most.

Shadow AI governance is an identity discovery problem before it is an AI policy problem. Agencies cannot govern tools they cannot enumerate, and they cannot enumerate tools that arrive through browser extensions, connectors, or delegated access outside procurement workflows. The practical implication is that identity governance must extend to all AI-connected access paths, including unsanctioned ones, because unmanaged identities create the same exposure route as unmanaged service accounts.

Orphaned access is the named failure mode hiding inside most modernization programmes. In consolidated and hybrid environments, permissions outlive role changes, contractor exits, and system migrations because lifecycle controls do not keep up with infrastructure change. This is where entitlement review, deprovisioning, and policy enforcement become one operating model across human users, machine accounts, and AI identities. Practitioners should treat access persistence as a structural governance defect, not an administrative delay.

Excess privilege is the common risk language across humans, machines, and AI agents, but the control response is not identical. Human access usually fails through role creep, NHI access through credential persistence, and AI agent access through runtime scope expansion. The article’s value is that it forces agencies to see those as variants of one governance problem. The implication is that access management teams need actor-specific controls under one policy model, not three disconnected programmes.

Identity blast radius is the right concept for state digital transformation. As agencies connect cloud services, analytics, and AI to the same data estate, the size of the damage from a single compromised identity becomes the decisive design variable. That is why zero-trust language alone is insufficient unless it is translated into concrete entitlement boundaries, lifecycle discipline, and continuous visibility. Practitioners should measure whether every new service reduces or expands blast radius before they approve it.

From our research:

• 92% of organisations expose NHIs to third parties, raising concerns about supply chain security, according to Ultimate Guide to NHIs.
• Only 20% have formal processes for offboarding and revoking API keys, and even fewer have procedures for rotating them, according to Ultimate Guide to NHIs , Lifecycle Processes for Managing NHIs.
• For a broader control map, use Top 10 NHI Issues to prioritise discovery, lifecycle, and privilege reduction across machine access.

What this signals

Identity-first transformation will increasingly be judged by whether agencies can see and govern every access path, not by how quickly they adopt new tools. In practice, that means identity teams must extend policy and lifecycle controls into AI-connected and cloud-connected workflows before those paths become normalised. NIST Cybersecurity Framework 2.0 remains the cleanest way to translate that discipline into governable functions.

Orphaned access will remain the most expensive symptom of modernization if offboarding stays manual. Our research shows 79% of organisations have experienced secrets leaks, and 77% of those incidents caused tangible damage, which is exactly why delayed revocation and weak ownership keep showing up as programme failures. Agencies should treat access cleanup as a delivery metric, not an administrative afterthought.

Identity blast radius is the concept teams should use to prioritise remediation. Once cloud services, AI tools, and legacy systems share the same data plane, the cost of a single excessive entitlement rises sharply. The practical question is no longer whether a system is modern enough, but whether its access model shrinks or expands the damage a compromised identity can cause.

For practitioners

• Map all AI-connected access paths Identify sanctioned and unsanctioned AI tools, extensions, and connectors that can reach internal data. Tie each path to an owner, a business purpose, and a revocation process so shadow AI does not become a hidden access layer.
• Unify lifecycle controls across identity types Apply joiner-mover-leaver and access review processes to human users, service accounts, and AI identities under the same governance standard. The goal is consistent revocation, not three separate workflows with different accountability.
• Reduce privilege before modernising platforms Baseline current entitlements in cloud and legacy systems, then remove excess access before expanding automation or AI use cases. Modernisation that inherits standing privilege simply moves risk into a faster environment.
• Track orphaned access as a migration risk During consolidation projects, require named owners for every legacy account and verify that deprovisioning completes before systems are retired. Orphaned accounts should block cutover until they are resolved.

Key takeaways

• State digital transformation now depends on identity governance that spans humans, machines, and AI agents, not siloed access tooling.
• Shadow AI, orphaned access, and privilege creep are the recurring failure patterns that turn modernisation into exposure.
• Agencies should treat lifecycle discipline and entitlement reduction as prerequisites for cloud and AI adoption, not cleanup tasks afterward.

Key terms

• Shadow AI: Unapproved or unmanaged AI tools, extensions, or agents that operate inside an environment without formal governance. In identity terms, shadow AI matters because it may carry delegated access, inherit permissions, or reach sensitive data without a clear owner or revocation path.
• Identity blast radius: The amount of damage a compromised identity can cause across connected systems, data, and services. For NHI and AI programmes, blast radius is shaped by privilege scope, credential persistence, and how widely one identity can move across cloud and legacy environments.
• Orphaned access: Access that remains active after the original owner, role, contractor relationship, or system context has changed. It is a lifecycle failure that increases exposure because permissions persist beyond accountability, especially in migration and consolidation programmes.
• Unified control plane: A single governance layer that applies consistent policy, visibility, and lifecycle controls across human identities, machine identities, and AI agents. The value is not centralisation for its own sake, but the ability to apply one access model across mixed estates.

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 building or maturing identity security in your organisation, it is worth exploring.

This post draws on content published by SailPoint: How SailPoint adaptive identity helps NASCIO’s top 10 priorities. Read the original.