Illumio Joins WWT's ARMOR Framework for AI Breach Containment

Illumio will be featured in World Wide Technology's AI Readiness Model for Operational Resilience, a vendor-neutral framework for assessing and managing AI risk, with the segmentation vendor's role centred on breach containment. The announcement, dated 11 November 2026, positions Illumio inside a model built to help organisations evaluate security problems tied to AI environments, specifically limiting how far an attack can spread after an initial compromise.

The stakes are straightforward. AI systems multiply the connections between applications, infrastructure, devices and external services, which increases the number of paths an attacker can use to move across an organisation's estate. Illumio says its products help security teams identify those paths and apply controls that reduce lateral movement — the spread of an attacker from one compromised system to others — across the systems supporting AI workloads.

What does the ARMOR framework actually cover?

WWT developed ARMOR together with NVIDIA and through work with the Texas A&M University System. The model organises AI security into seven domains:

• Governance, Risk and Compliance
• Model Protection
• Secure AI Operations
• Infrastructure Security
• Identity Security
• Data Security
• the Secure Software Development Lifecycle

Within that structure, Illumio sits in the infrastructure and operational resilience layers. The emphasis there is containing a breach rather than assuming preventative measures will stop every attack. That framing reflects a wider industry shift toward resilience planning as organisations connect AI tools to core business systems — accepting that some intrusions will succeed and planning around what happens next.

Under the framework, Illumio says it helps organisations map communications across workloads, devices and internet connections to identify high-risk exposure around AI systems. Customers can also apply segmentation policies — controls that divide a network into isolated zones — across data centres, cloud environments, containers, endpoints and operational technology.

The framework also addresses public AI services. That includes identifying connections to public large language model services, monitoring AI agent processes and the systems they contact, and measuring whether observed traffic matches policy requirements and control standards. This matters because corporate AI estates increasingly mix internal infrastructure with external model providers, giving security teams a broader perimeter to police than traditional enterprise networks.

One technical detail stands out. Illumio says its segmentation measures can extend to performance-sensitive AI infrastructure through NVIDIA BlueField-3 DPUs — data processing units that offload networking and security tasks from the main processor — allowing controls to be enforced beyond the host itself. For AI clusters where latency matters, that distinction determines whether security controls are practical or skipped entirely.

What did the executives say?

Todd Palmer, Senior Vice President of Global Partner Sales and Alliances at Illumio, said the company's inclusion in ARMOR reflects a shift in how companies must think about cyber defence in AI deployments.

"AI isn't just adding more technology to the enterprise; it's creating systems that are more connected and increasingly autonomous. That makes controlling how far and how fast an attack can move more important than ever. ARMOR gives organisations a framework for thinking about that risk, and Illumio brings the containment needed to limit its impact," Palmer said.

Chris Konrad, Vice President of Global Cyber at World Wide Technology, outlined WWT's view of the need for broader controls around AI systems.

"Securing AI requires more than protecting a model in isolation. Organisations need control across the infrastructure, identities, applications, and data that allow AI to operate. With its differentiated ability to see and contain lateral movement across complex environments, Illumio contributes to ARMOR's broader approach to helping organisations build secure, resilient AI environments," Konrad said.

Both statements point to the same premise: AI security cannot be reduced to protecting a single model, because the risk lies in the connections the model creates.

Why is containment the focus?

The announcement highlights a growing concern among corporate security teams: AI deployments can expand attack surfaces by linking models, data stores, cloud platforms, endpoints and third-party services. In that context, vendors are increasingly framing cyber resilience around what happens after an attacker gains initial access — including whether internal communications can be restricted quickly enough to prevent wider disruption.

Illumio said its platform combines visibility and segmentation through its Illumio Insights and Illumio Segmentation products. That combination lets security teams understand risk in connected environments and isolate exposed systems or suspicious pathways when they are found.

For WWT, the addition broadens the set of suppliers represented in ARMOR as customers look for ways to evaluate AI security across multiple layers rather than through a single product category. The framework's vendor-agnostic positioning suggests it is intended to serve as a reference model for organisations still determining how to structure AI governance and operational safeguards.

What is the broader picture?

ARMOR groups its safeguards across governance, model, infrastructure, identity, data and software development controls, while Illumio's contribution remains centred on limiting lateral movement across the systems that support AI operations. The division of labour is deliberate. WWT supplies the assessment structure across seven domains; Illumio supplies the containment capability within two of them.

The deal also illustrates how channel partners and systems integrators are becoming aggregation points for AI security. Rather than selling one product, WWT is assembling a portfolio mapped to a published framework, developed with NVIDIA and validated through work with the Texas A&M University System. For vendors, inclusion in such a model is a distribution channel; for buyers, it is a way to compare capabilities across a fragmented market.

The forward question is adoption. If ARMOR becomes a widely used reference model, placement within its seven domains will shape how enterprises shortlist security vendors for AI environments — and containment, the layer Illumio occupies, will be measured by how quickly organisations can restrict internal communications during a real incident.