From AI Discovery to Attack Surface Mapping: Announcing the Wiz + Pillar Partnership

Dor Sarig

and

March 23, 2026

min read

We're announcing today our partnership with Wiz. By combining Wiz's AI discovery in the cloud with RedGraph - Pillar's agentic AI red teaming engine, joint customers can now map their AI attack surface, validate the agentic runtime risks, and see a unified view of AI-specific findings in the Wiz Security Graph, all in one connected workflow.

As AI agents grow more autonomous and enterprises embed models deeper into cloud workloads, the AI attack surface continues to expand. Platforms like Wiz provide deep visibility into where AI systems run, how they’re connected, and whether they’re exposed. Building on this foundation, organizations can further validate how those systems behave in practice, testing for prompt leakage, adversarial manipulation, and potential data exfiltration through agent tools and interactions. Together, this layered approach ensures AI systems are not only visible and monitored for exposure, but continuously tested and secured against real-world threats.

The Pillar + Wiz integration closes that gap.

Wiz discovers AI workloads across your cloud environment, maps the services, identities, and data they’re connected to, and surfaces any security risks on the Wiz Security Graph.

Pillar Security picks up from there, testing those AI endpoints with RedGraph, our autonomous red-teaming engine, and feeding the results back into the Wiz Security Graph as enriched, evidence-backed risk data.

What Joint Customers Get

Most organizations have limited visibility into what their AI systems actually do, what data they access, and what is the potential blast radius of a successful attack. Cloud security gives you the infrastructure picture. AI security gives you the agentic-layer picture.

The Wiz + Pillar integration bridges these 2 worlds:

Complete AI asset inventory. Wiz discovers every AI workload in your cloud, from managed services to self-hosted models to agentic endpoints, and identifies which ones are exposed. No blind spots.

Automated AI attack surface mapping. Pillar's RedGraph engine autonomously probes discovered AI endpoints for vulnerabilities: prompt injection, system prompt extraction, tool discovery, safety bypasses, and data exfiltration risks. Every test produces evidence, including session recordings of the adversarial interactions.

Enriched risk data in the Wiz Security Graph. RedGraph results feed directly back into Wiz, so security teams see AI-specific risk data alongside cloud context. You can prioritize based on real, validated exposure rather than theoretical risk.

Continuous validation. The integration runs on a scheduled cadence, so your AI attack surface gets tested regularly, not once.

Why We're Partnering

"AI agents are now your biggest attack surface - and they are being built and deployed without your security team in the room. Wiz and Pillar change that: full cloud context meets full agentic depth, giving CISOs one unified view of their real agentic AI risks, from deployment to runtime." Dor Sarig, Co-Founder & CEO, Pillar Security

"As AI accelerates, security is built together. We’re excited to welcome Pillar to the Wiz Integration Network as a new AI Security partner. By ingesting Pillar’s AI findings into Wiz, joint customers gain deeper visibility into AI-specific risks and a clearer understanding of their AI attack surface." Oron Noah, VP of Product, Extensibility & Partnerships, Wiz

A New WIN Category, and Pillar Covers All of It

Wiz recently launched a dedicated AI Security category within the Wiz Integration Network (WIN), extending the platform’s native AI and cloud security capabilities through an open ecosystem of specialized partners. This approach brings together adversarial testing, runtime guardrails, agentic AI security, and governance into a connected layer of intelligence, where partners build on Wiz’s deep context to secure AI workloads together. Pillar joins WIN as part of that new category.

Pillar brings all four of the WIN AI security capabilities in one platform:

AI Discovery & Posture Management: inventory and posture assess every AI component
Red Teaming & Risk Detection: autonomous adversarial testing and attack surface mapping at scale
Runtime Guardrails: adaptive protection layer that blocks threats and policy violations in real time
Governance & Compliance: continuous monitoring and benchmarking against AI security frameworks and standards

The Wiz integration focuses on the first two: discovery and red teaming. Wiz provides the cloud security context and surfaces exposed AI endpoints. Pillar validates the risk with RedGraph and sends evidence back. For customers who want to go further, Pillar's runtime guardrails and governance capabilities can be configured independently to act on those same results.

Getting Started

If you're already a Wiz customer, enabling the Pillar integration takes minutes. Connect your Wiz credentials within Pillar's platform, configure your scan schedule, and RedGraph begins assessing your AI endpoints automatically.

For organizations new to Pillar, our team can walk you through a live demo showing the end-to-end flow, from Wiz discovery to RedGraph results enriching the Security Graph.

Request a demo →

FAQs

How does the Wiz and Pillar Security integration map an AI attack surface?

Wiz discovers every AI workload across the cloud environment — managed services, self-hosted models, and agentic endpoints — and identifies exposed assets. Pillar's RedGraph engine then autonomously probes those discovered endpoints for prompt injection, system prompt extraction, tool discovery, safety bypasses, and data exfiltration risks, feeding validated, evidence-backed results directly into the Wiz Security Graph.

What is RedGraph and what vulnerabilities does it test for in AI agents?

RedGraph is Pillar Security's autonomous red-teaming engine designed specifically for agentic AI systems. It probes AI endpoints for prompt injection, system prompt extraction, tool discovery, safety bypasses, and data exfiltration risks. Every test produces evidence including session recordings of adversarial interactions, giving security teams validated exposure data rather than theoretical risk assessments.

Why do enterprises need both cloud security and AI-layer security to protect AI agents?

Cloud security provides the infrastructure picture — where AI systems run, how they're connected, and whether they're exposed. AI security adds the agentic-layer picture — how those systems actually behave, what data they access, and the potential blast radius of a successful attack. Neither layer alone is sufficient; the Wiz and Pillar integration bridges both into one connected workflow.

What four AI security capabilities does Pillar Security bring to the Wiz Integration Network?

Pillar covers all four AI security capabilities in Wiz's new WIN AI Security category: AI Discovery and Posture Management, Red Teaming and Risk Detection via autonomous adversarial testing, Runtime Guardrails that block threats and policy violations in real time, and Governance and Compliance through continuous monitoring against AI security frameworks. The Wiz integration specifically activates the first two.

How do joint Wiz and Pillar Security customers get started with continuous AI attack surface validation?

Existing Wiz customers can enable the Pillar integration in minutes by connecting Wiz credentials within Pillar's platform and configuring a scan schedule. RedGraph then automatically assesses AI endpoints on a continuous, scheduled cadence — meaning the AI attack surface is tested regularly rather than as a one-time assessment — with results enriching the Wiz Security Graph as they arrive.