Commvault & Zscaler partner on enterprise AI security

WorkAI.TV Editorial Desk
3 Min Read

Share with your CISO

Commvault and Zscaler are betting that enterprise AI security breaks into two distinct problems, detection and recovery, and that neither vendor can own both alone. Their announced partnership combines Zscaler’s AI application discovery and policy enforcement with Commvault’s data and workload recovery products. The joint focus covers three areas: visibility into sanctioned and shadow AI use, runtime governance for AI agents, and clean-state restoration after an AI-driven incident disrupts data, applications, or identity systems.

What this means for your business

Shadow AI, meaning AI tools employees and developers adopt without IT approval, is the problem that makes this partnership legible. If your organization has moved AI agents into production workflows where they read, write, or act on business data, you already have an exposure you may not be measuring. The CISO whose team is still treating AI risk as a subset of standard SaaS governance is the one this story is about. The CISO who has already built a recovery runbook specifically for AI-driven incidents is the one who isn’t surprised by it.

The structure of this deal reflects a gap the market is quietly admitting. Prevention-only security architectures, which dominated the zero-trust conversation for the past decade, assume a human actor making a discrete decision. AI agents break that model because they act continuously, at speed, across data boundaries, and a single misconfigured permission or poisoned prompt can propagate damage before any alert fires. Recovery capability, the ability to restore a known-good state after an AI system corrupts or exfiltrates data, stops being a DR (disaster recovery) afterthought and becomes a live security control. Commvault’s positioning here is strategically sharp, even if the partnership announcement itself is light on the integration specifics that would let a buyer actually evaluate it.

Zscaler’s framing, coming from a vendor whose network-layer visibility is the commercial hook, tilts naturally toward discovery and policy as the primary value. That’s worth noting because the harder unsolved problem isn’t knowing which AI tools your employees are using; it’s knowing what your AI agents are doing autonomously at the data layer, and whether your recovery tooling has ever actually been tested against an AI-caused incident rather than a ransomware scenario. If your current incident response playbooks don’t include a step for “AI agent took an action we didn’t authorize,” that’s the gap this partnership is pointing at. I’d revisit this partnership as genuinely differentiated only once Commvault and Zscaler publish a documented, tested integration with a named recovery workflow rather than a joint press release.

Based on reporting from Commvault & Zscaler partner on enterprise AI security, originally published 2026-11-16 19:05:00.

TAGGED:
Share This Article