Zscaler Expands Project AI-Guardian with OpenAI, AWS, and Google Cloud to Secure Enterprise AI

• AWS
• CoreWeave
• Databricks
• Deep Cogito
• Equinix
• Glean
• Google Cloud
• OpenAI
• Saviynt

New Global System Integrator Partners

Two additional GSIs now join the original founding cohort of Cognizant, EY, HCLTech, Infosys, TCS, and Wipro :

• Coforge
• NTT DATA

How Partner Technologies Integrate with the Zero Trust Exchange

The core architectural principle behind the expansion is embedding Zscaler’s security signals directly into partner platforms—and vice versa—through the Zero Trust Exchange. This isn't a loose collection of APIs; it is a tightly coupled signal-sharing framework where visibility and enforcement flow in both directions .

Two Zscaler AI services sit at the heart of this integration :

• AI Access Graph: Continuously maps how identities, applications, agents, and data connect to sanctioned and unsanctioned AI services. Partners receive this live identity and access context to enhance their own policy engines.
• AI Attack Surface and Risk Modeling: Discovers, inventories, and quantifies AI-related risk exposure across the enterprise. Risk scores and asset maps are shared with partner platforms so security teams can prioritize threats in the tools they already use.

Because every enterprise AI interaction is brokered inline through the Zero Trust Exchange, enforcement happens in real time. An insight generated in Databricks can drive a zero-trust policy action within AWS. A risk model update from Zscaler can inform Glean's access permissions. The goal is a consistent, unified control plane that spares security teams the labor of manually correlating alerts across disconnected dashboards .

The Challenge: No Single Vendor Can Secure AI Alone

Zscaler explicitly framed this expansion as a remedy for a market-wide failure. As organizations accelerate AI adoption, security teams confront a sprawling attack surface spanning multiple AI models, cloud environments, agent frameworks, and SaaS applications. Without interoperable tools, each new AI service creates a fresh blind spot.

In the words of the official press release, “No single vendor can secure the entire AI estate on its own,” and the industry's default approach—manual integration of fragmented point solutions—simply does not scale for AI . Project AI-Guardian is built around the principle that a comprehensive AI security platform must share signals, identity context, and enforcement with both leading and existing enterprise technologies .

Executive Perspective: Security as an Ecosystem Effort

Dhawal Sharma, Zscaler’s EVP of AI Security and Strategic Initiatives, underscored the collaborative philosophy at the center of the announcement:

“AI is creating enormous opportunities for organizations, but it is also reshaping the threat across the security and governance landscape. Securing AI is an ecosystem effort. With the expansion of Project AI-Guardian through our technology alliance partners, Zscaler is helping customers extend zero trust across enterprise AI interactions so they can adopt AI faster while maintaining the visibility, control, and data protection they need to innovate securely.”

Regarding the technology alliance partners, the available sourcing confirms OpenAI joined the initiative as a partner, and CoreWeave’s participation is noted in connection with “securing AI infrastructure and agentic workloads at scale” through shared commitments . However, the complete verbatim executive quotes from OpenAI and CoreWeave leadership are not included in the provided source excerpts .

From GSIs to a Full Ecosystem

The original Project AI-Guardian launch in May 2026 focused on combining Zscaler’s AI security platform with the consulting and deployment expertise of global system integrators to help enterprises transition into the “agentic world” securely . The Zenith Live 2026 expansion layers on native technology integrations, moving the initiative from advisory and implementation acceleration into a live, interoperable security fabric that spans the operational tools enterprises already use every day.