How to Secure GenAI Usage in the Enterprise
Zscaler, Thursday, November 5th, 2026
Zscaler recommends a zero trust approach to securing enterprise GenAI use, with inline inspection, context-aware DLP and browser isolation.
Zscaler's ThreatLabz 2026 AI Security Report found enterprise AI and ML transactions rose 91% in 2025 to 989.3 billion, with 18,033 TB of corporate data sent to AI apps and over 410 million DLP violations tied to ChatGPT.
Enterprises blocked 39% of AI transactions, which Zscaler says can push employees toward unsanctioned tools, while unrestricted access risks intellectual property.
It argues static controls such as URL filters miss risk in prompt content, so continuous zero trust controls should verify user, device, app and content for every session.
Recommended measures include inline SSL inspection with context-aware DLP to allow, redact or block prompts and cloud browser isolation for GenAI sessions.