Identity-driven microsegmentation is overtaking IP-based zero trust

At a glance

What this is: This is Gartner’s view that network microsegmentation must move beyond IP boundaries toward identity and context, especially as AI agents expand the number of workload identities in play.

Why it matters: It matters because IAM, PAM, and NHI teams now need segmentation logic that understands who or what is acting, not just where traffic originates.

By the numbers:

• 92% of organisations expose NHIs to third parties, raising concerns about supply chain security.
• Only 5.7% of organisations have full visibility into their service accounts.
• When AWS credentials are exposed publicly, attackers attempt access within an average of 17 minutes, and as quickly as 9 minutes in some cases.

👉 Read Zero Networks' report on Gartner's identity-led microsegmentation findings

Context

Network microsegmentation is the practice of limiting east-west movement inside an environment, but IP-based policy alone cannot distinguish between legitimate workload communication and identity-driven abuse. Gartner’s report argues that the control plane has shifted toward identity and context because AI agents, workloads, and service identities now move faster than traditional network assumptions.

For IAM and NHI practitioners, the important point is not whether segmentation exists, but whether it is keyed to identities that can be governed across lifecycle, privilege, and runtime context. That makes this topic relevant to zero trust programmes, workload access control, and any security model that still treats network location as the main signal.

Key questions

Q: How should security teams implement identity-aware microsegmentation in hybrid environments?

A: Start by tying segmentation policy to authenticated workload, service account, and agent identity, then map each identity to the smallest set of allowed destinations. Use runtime context to tighten exceptions, and review any rule that depends only on IP address or subnet. The goal is to make lateral movement fail by default, not to add another policy layer on top of permissive network design.

Q: Why do AI agents make static network segmentation less effective?

A: AI agents can create, chain, and repeat internal actions faster than teams can review network-based exceptions. They also blur the line between normal service-to-service traffic and delegated activity, which makes IP boundaries too coarse for reliable control. Static segmentation cannot express task scope well enough, so identity and context must become part of the policy decision.

Q: What breaks when segmentation ignores service and workload identity?

A: When segmentation ignores identity, any caller that reaches an allowed network location may inherit access that was meant for a specific workload or role. That creates a broad blast radius for compromised credentials, shared tokens, and misrouted automation. The failure mode is not just unauthorised access, but movement across systems that should have remained isolated.

Q: Who is accountable when microsegmentation fails to contain lateral movement?

A: Accountability usually spans IAM, network security, cloud security, and platform engineering because the failure sits at the boundary between identity governance and traffic enforcement. Teams should define owners for policy design, exception approval, and runtime monitoring before an incident occurs. Without that clarity, the control exists in theory but not in operational practice.

Technical breakdown

Why IP-based microsegmentation breaks down in identity-heavy environments

Traditional microsegmentation uses subnets, ports, and labels to decide what can talk to what. That works poorly when workloads are ephemeral, identities are shared, and AI agents can invoke tools or services across changing execution paths. The problem is not just scale, but semantic blindness: an IP address tells you where traffic came from, not whether the caller is an approved workload, a compromised service account, or an agent acting outside its intended scope. In identity-heavy environments, the segmentation policy needs identity context, not just network topology.

Practical implication: align segmentation policy with workload and service identity rather than relying on static network objects.

Identity as the control plane for zero trust segmentation

A zero trust segmentation model treats identity as the control plane because access decisions should be driven by authenticated workload, service, or agent identity plus runtime context. That means policy can reflect trust boundaries, task scope, and privilege rather than coarse IP zones. For NHI governance, this is especially relevant where service accounts, tokens, and AI agents can initiate calls on behalf of applications. The architectural shift is from a passive network filter to a decision layer that continuously evaluates who the caller is and what it is allowed to do.

Practical implication: integrate identity signals into policy enforcement points so lateral movement is constrained by authenticated context.

Agentic AI changes what blast-radius control has to do

Agentic AI increases the speed and ambiguity of internal traffic because autonomous or semi-autonomous systems can chain actions across multiple tools and services. In that setting, blast-radius control is no longer just a containment exercise after detection. It becomes a live governance function that limits how far an agent, workload, or compromised identity can move before human review. That is why the report’s emphasis on active defence matters: the control must operate during the action, not only after the alert.

Practical implication: design segmentation and privilege boundaries to stop cross-service chaining by AI agents and compromised NHIs.

Threat narrative

Attacker objective: The objective is to expand a single foothold into broad internal access by exploiting segmentation models that cannot reliably distinguish identity from network position.

1. Entry begins when an attacker or malicious agent obtains a foothold inside the environment and reaches internal services that still trust network location more than identity.
2. Escalation occurs when that foothold is able to reuse workload access paths or service identities to move between services that should have been isolated by policy.
3. Impact follows when lateral movement is no longer constrained, allowing broader compromise of internal systems, data flows, or operational workloads.

NHI Mgmt Group analysis

Identity-led segmentation is becoming the new baseline for zero trust. Gartner’s framing reflects a real governance shift: network controls now have to understand the identity of workloads, service accounts, and AI agents, not just their addresses. Static zones cannot keep pace with ephemeral compute and delegated runtime activity. Practitioners should treat identity-aware policy as the design centre of segmentation, not an optional enhancement.

AI agent proliferation turns microsegmentation into an NHI governance problem. As agents and automation layers grow, the number of non-human actors capable of initiating internal calls increases sharply. That expands the surface area for standing access, inherited trust, and lateral movement through over-permissioned identities. The control question is no longer simply which subnet a workload sits in, but which identities can legitimately traverse which boundaries. Practitioners should govern segmentation as part of NHI lifecycle and privilege management.

Blast-radius control is becoming an active defence function, not a post-breach containment task. The report’s most useful idea is that segmentation should reduce attacker reach before humans intervene. That aligns with modern identity security thinking, where delay between compromise and containment is often the real failure. If policy cannot act in runtime, it is too late to stop movement. Practitioners should design for real-time containment across identity paths, not just after-the-fact forensics.

Agentless innovation only matters if enforcement is still policy-complete. The operational appeal of agentless segmentation is deployment speed, but speed is not the security outcome. What matters is whether enforcement remains precise enough to preserve least privilege across servers, endpoints, and OT assets. Where the control plane cannot see the identity layer, policy drift becomes inevitable. Practitioners should verify that convenience does not degrade the fidelity of access control.

Zero trust programmes will fail if they treat identity and network segmentation as separate disciplines. Gartner’s report reinforces what identity teams have been seeing for years: access control, workload identity, and east-west segmentation are converging. The organisational implication is that IAM, PAM, cloud security, and network teams need shared policy language. Practitioners should align governance so identity decisions flow into segmentation decisions, not around them.

What this signals

Identity-aware segmentation is moving from design preference to governance requirement. Programmes that still separate network policy from identity policy will struggle to contain delegated workload activity, especially where AI agents can call services in chained sequences. The practical signal is that segmentation reviews now belong alongside access reviews, not outside them.

One useful concept here is identity-bound blast radius: the maximum internal reach a workload, service account, or agent can achieve before policy stops it. If that boundary is unclear, a single compromise can still become a multi-system event. Practitioners should measure whether their enforcement points actually constrain that radius in live traffic, not only in policy documents.

For teams using zero trust language, the test is whether enforcement can act on authenticated context quickly enough to matter. Where policy changes lag behind runtime behaviour, lateral movement wins. That makes workload identity telemetry, exception governance, and cross-domain ownership central to the next phase of segmentation programmes.

For practitioners

• Map segmentation policy to workload identity Replace IP-only allowlists with rules that bind traffic permissions to authenticated workload, service account, or agent identity, then review where identity context is missing from enforcement.
• Define blast-radius limits for NHIs and agents Set explicit cross-service boundaries for service accounts, tokens, and AI agents so a compromise cannot pivot across unrelated application tiers or operational zones.
• Bring segmentation into identity governance Add microsegmentation decisions to IAM and PAM governance so access reviews include runtime communication paths, not only entitlements and privilege assignments.
• Validate policy completeness across OT and cloud assets Test whether the same identity-aware enforcement model works across servers, laptops, cloud workloads, and OT assets, then close gaps where traffic is still governed by network location alone.

Key takeaways

• Gartner’s core argument is that network segmentation must move beyond IP boundaries and into identity-aware control.
• The scale issue is not just workloads but the growing number of NHIs and AI agents that can traverse internal services.
• Practitioners should treat blast-radius containment as a live governance function shared across IAM, PAM, cloud, and network teams.

Key terms

• Microsegmentation: Microsegmentation is the practice of dividing internal networks into small policy zones so only approved traffic can move between them. In modern environments, it should be enforced with workload and identity context, not just IP addresses, so that trust follows the caller rather than the subnet.
• Identity-Led Segmentation: Identity-led segmentation uses authenticated workload, service account, or agent identity as the main basis for traffic decisions. It is designed to limit lateral movement by making policy depend on who or what is calling, which is essential when workloads are ephemeral and network location is no longer a reliable trust signal.
• Blast Radius: Blast radius is the amount of damage or reach an attacker can achieve after compromising one account, workload, or system. In identity security, it is a practical measure of how far a stolen credential or abused agent can move before controls stop the chain of access.

What's in the full report

Zero Networks' full report covers the operational detail this post intentionally leaves for the source:

• How Gartner frames identity-led segmentation across AI agents, workloads, and zero trust architecture
• The report’s discussion of agentless microsegmentation and what that changes operationally for deployment
• The specific implications for shrinking blast radius before human detection and response
• How the report positions network segmentation in a broader security roadmap

👉 Zero Networks' full report covers the Gartner context, segmentation framing, and roadmap implications in more detail.

Deepen your knowledge

NHI Foundation Level course, the industry's only accredited NHI security programme, covers NHI governance, workload identity, secrets management, and identity lifecycle fundamentals. It is designed for practitioners aligning identity control with broader security architecture and operational risk.