On September 9, 2026, Zscaler announced Zscaler Agentic SOC, a security operations model built to detect and automatically contain AI-driven threats, according to its press release.

The company says the offering pairs its own telemetry, claiming 750 billion daily zero-trust transactions, with a decoy mesh network, native zero-trust controls, third-party integrations, and AI agents. Zscaler also says it has partnered with Anthropic and OpenAI to integrate their AI models into the pipeline. Neither AI lab independently confirmed the partnerships.

Zscaler’s ThreatLabz team reports growing use of evasive tactics such as abuse of trusted sites and legitimate remote-management tools.

For enterprise and federal SOC leaders, the pitch lands on a real gap: attackers automate while analyst triage stays manual.

However, Zscaler discloses no general availability date, pricing, customer count, or independently benchmarked performance figures, so efficacy claims remain vendor-asserted. And routing security telemetry through external AI models will raise data-handling and vendor-risk questions that regulated buyers should resolve before adoption.

Source: Zscaler