Commvault and Zscaler Plan Enterprise AI Agent Recovery

By Nathan Brooks

Commvault and Zscaler announced a strategic partnership on November 10, 2026, combining AI-use discovery and policy enforcement with recovery for enterprise data, applications and identities. Commvault said the companies plan to help enterprises govern production AI agents and roll back harmful or unintended activity.

Commvault and Zscaler said the partnership addresses the operational gap between controlling an agent and repairing what it changes. In the companies’ plan, Zscaler Security for AI identifies enterprise use of AI applications and enforces policies for users and AI agents.

Commvault’s AI and cyber-resilience technology restores data, applications and identities after attacks, disruptions or outages. Commvault and Zscaler said those functions create a proposed path from detection through recovery.

What does this change for security and IT teams?

The stakes rise when experimental agents become production systems with direct access to business data. Commvault and Zscaler said security and IT teams must identify which AI tools employees and developers use, what data those tools share, which activities to permit or block and whether unwanted changes can be reversed quickly.

Commvault and Zscaler connect preventive controls with a recovery path for changes that slip past them. Zscaler would supply visibility and enforcement, while Commvault would address restoration after an incident. For practitioners, that division matters, but the work remains a planned collaboration rather than a generally available joint product.

Commvault Chief Technology and AI Officer Pranay Ahlawat said agents ‘do not behave deterministically’ and that a single agent decision ‘can trigger changes across data, applications and infrastructure on multiple platforms.’ Commvault said such an unintended action can propagate at machine speed, affecting data integrity, application availability and business operations.

Three planned workstreams

Commvault and Zscaler named AI discovery, runtime governance, and agent detection and recovery as the partnership’s use cases. Commvault described them as planned collaboration areas rather than a generally available joint product. The companies outlined three workstreams:

• AI discovery: Expand enterprise-wide discovery and maintain continuous visibility into active AI applications, models and tools.
• Runtime governance: Flag unsanctioned AI activity and prevent sensitive-data exposure.
• Agent detection and recovery: Pair real-time protection for autonomous agent operations with rapid, clean recovery after a security event or risky AI action.

AgentRecover’s recovery model

Commvault provided a more detailed view of its recovery approach in a separate AgentRecover announcement. The new capabilities support Amazon Bedrock AgentCore and Microsoft Foundry, platforms that provide services for building, deploying and operating production agents. AgentRecover is designed to audit agent activity, restore agents and recover affected data and systems across environments.

For supported platforms, AgentRecover is designed to discover agents and map the data, applications, identities, configurations and infrastructure with which they interact. Commvault said the dependency map is intended to reveal protection gaps. Its auditing functions are designed to establish what an agent did and what it may have affected across platforms, helping teams prioritize recovery work.

Commvault said recovery covers both the agent and the affected resources while concentrating on what the agent changed. Customers can also use AgentRecover with Commvault’s identity recovery and change auditing to examine agents’ actions, the data they changed and the identities through which they acted.

Commvault is targeting early access for Microsoft Foundry in November 2026 and for Amazon Bedrock AgentCore in December 2026. Those targets apply to AgentRecover’s rollout; the Zscaler announcement does not make them availability dates for the partnership’s planned work.

For joint customers, the central test is whether the companies can turn the planned use cases into working discovery, governance and agent-recovery capabilities. Until that happens, the partnership’s immediate significance is its division of labor, not a shipping joint product.