Zscaler AI data protection and the SASE proxy competition

AI threats and data visibility

Zscaler’s October AI Data Protection launch targets the massive expansion of AI-driven threats. Enterprise AI usage rose 91% last year, and data transfers to AI/ML applications reached 18,033 terabytes in 2025. Zscaler red team testing found the median time to the first critical failure at 16 minutes. One-shot attacks that use a single prompt produced a 60% failure rate. It manages risks from 3,400 different AI apps used by employees. I find the focus on shadow AI compliance necessary since workers use these tools without IT approval. The suite provides AI asset discovery, Zero Trust controls, and risk assessment for agentic AI that operates at speeds exceeding humans. This suite also includes real-time Email DLP for Microsoft 365 and Google Gmail. Zscaler analyzes and harnesses knowledge from over 500 trillion data points daily across users, devices, networks, and applications. The platform protects structured and unstructured data in-motion, at-rest, and in-use across all inline channels including Web, SaaS, Email, BYOD, and private applications. This includes data in SaaS, IaaS/PaaS environments, endpoints, and on-premises network shares. The platform uses AI to discover and classify sensitive information across all data-at-rest locations including endpoint, SaaS, and public cloud infrastructure to reduce the complexity of managing multiple legacy point products.

Zscaler versus Netskope

Netskope holds the top spot in the 2026 Gartner SASE Platforms ranking, while Zscaler recently secured its first year as a SASE leader. Zscaler leads in ZTNA scale and VPN replacement for large enterprises. Netskope maintains an advantage in CASB and data governance depth. Netskope’s CASB provides visibility into 70,000 plus SaaS apps and includes 100 plus app activity controls. Zscaler splits its offering into Zscaler Internet Access for web traffic and Zscaler Private Access for zero-trust application access. Both companies require endpoint clients like the Zscaler Client Connector or the Netskope Security Client to reach private applications. Both vendors use a proxy architecture that routes traffic through a middle hop, which adds latency for distributed teams. This routing creates delays when users access resources from locations far from an enforcement node. If you manage a team in China, both vendors face known routing difficulties. I would skip the Zscaler platform if your team needs the deepest possible CASB granularity or SaaS posture scanning. Netskope is a consistent Magic Quadrant Leader and provides ZTNA, DLP, and a tight round-trip SLA for TLS inspection.

Scale and the final verdict

Zscaler processes over 500 billion daily transactions and secures 40% of Fortune 500 companies. The platform includes natively integrated Data Security Posture Management to discover and protect sensitive data in AWS and Microsoft Azure. It also identifies risks in prompts and AI app usage. The expansion of AI Auto Data Discovery now covers endpoints, SaaS, and public cloud infrastructure. Zscaler ZIA starts in the range of 8 to 15 dollars per user per month, while Private Access is licensed separately. The ZDX module provides visibility into application performance and device health. Zscaler’s Risk360 delivers automated cyber risk quantification via real-time data. The platform provides 99.999% uptime and maintains a customer retention rate of 95%. For 20,000-seat deployments, Zscaler provides the necessary scale for ZTNA, though you already know that large scale requires significant engineering headcount. Can a proxy architecture ever fully eliminate the latency issues for a truly global, mobile workforce? You should choose Zscaler for ZTNA at enterprise scale and Netskope if your priority is cloud application data governance. Zscaler’s AI Security Suite introduces tools for AI asset discovery, Zero Trust controls, and AI risk assessment.