Commvault will let IT and security teams discover, protect, and recover ChatGPT and agent configurations plus other AI data stored in Anthropic's Claude Cowork and Claude Code, and in OpenClaw. The capabilities sit inside AgentRecover, Commvault's recoverability portfolio for enterprise AI environments.

The target is the endpoint. Employees run agentic AI tools on Windows and macOS PCs and laptops, and when a device is lost, damaged, or replaced, organizations risk losing more than files: agent configurations, accumulated AI context, and workflows built inside those tools can vanish with the hardware. Commvault's pitch is that it can restore that agent metadata.

Pranay Ahlawat, Commvault's Chief Technology and AI Officer, framed the visibility gap bluntly: "Most organisations cannot tell you which AI agents are running on their employees' machines, let alone recover the work inside them. We give IT and Security visibility into where these AI tools are installed, and a recovery path for the context employees have built inside them."

The product mechanics split along the IT/security line. IT teams can identify instances of supported agentic AI tools installed across managed devices, protect agent configurations stored locally, and recover that protected data after a disruption. Security teams get a centralized view of supported AI tools across managed endpoints — so if an agentic tool landed outside approved IT processes, they can see where it runs and investigate.

Availability is staged. Discovery of OpenAI's ChatGPT desktop app ships today. Discovery, protection, and recovery for supported local data tied to Claude Cowork, Claude Code, and OpenClaw reach general availability in the coming months. Customers get the features through Commvault's existing endpoint protection offerings — no separate SKU, no additional purchase. Support for further AI tools is on the roadmap for later releases.

A crowded field, with different scopes

Commvault is not alone in targeting agent data. Three competitors have shipped adjacent products, and the differences in scope matter to buyers.

Cohesity's Agent Resilience (CAR), a capability within Cohesity Data Cloud, protects agent memory and configuration, and supports Amazon Bedrock AgentCore and Bedrock Agents. It does not cover the Anthropic agents or OpenClaw — which makes Commvault's Anthropic and OpenClaw coverage a direct differentiator today.

Druva offers Claude Backup within its AI Resilience / Endpoints line. It backs up locally stored Claude Code and Cowork data on Windows and macOS: project files, sessions, tasks, prompts, notes, skills, plugins, connectors, settings, and configuration. Snapshots land in Druva's air-gapped cloud, and administrators manage policy centrally. One boundary: Claude.ai cloud chats are out of scope because they do not live on the device.

Rubrik's Rubrik Agent Cloud covers Claude Code and Claude Cowork with a heavier governance bent. It inventories agents, offers Agent Rewind to reverse unintended actions, keeps immutable off-repo snapshots of GitHub and Azure DevOps, and backs up and version-tracks agent configuration — system prompts, tool permissions, skills, settings — at organization, repository, and user levels, with drift detection. That positions Rubrik closer to agent governance and codebase recovery than laptop file backup, while Commvault anchors itself in endpoint recoverability.

Veeam, which has recently emphasized agentic AI cyber-threats, will likely follow Cohesity, Commvault, Druva, and Rubrik into this segment.

What's being protected

Anthropic's Claude Cowork is a knowledge-work agent available on macOS and Windows that can organize folders, turn notes and screenshots into spreadsheets, or draft reports from scattered sources. Claude Code is an agentic coding tool: it reads a codebase, edits files, runs commands, and works with development tools to build features, fix bugs, and automate tasks rather than only suggesting the next line of code.

OpenClaw is a free, open-source autonomous general-purpose agent that runs on the user's own system. It connects a large language model to messaging apps and local tools so it can act on tasks rather than only reply in a chat window. Users reach it through WhatsApp, Telegram, Discord, Slack, Signal, iMessage, other channels, a web UI, a CLI, or companion apps. Session state and memory stay on local hardware. Models are plug-ins: Claude, GPT, Gemini, DeepSeek, OpenRouter, or local models via Ollama and similar tools.

Separate AWS milestone

Commvault also announced it has achieved the AWS Data Competency in the Relational Database category, citing cyber-resilient data protection for relational database workloads on AWS. The designation covers helping customers protect, recover, and manage business-critical data across Amazon Relational Database Service (RDS) environments.

The competitive question now is speed. Commvault ships ChatGPT discovery today but locks full Claude Cowork, Claude Code, and OpenClaw protection behind a "coming months" window — a gap during which Rubrik and Druva, both already shipping Claude coverage, can press their head start in a market where agent sprawl on endpoints is only expanding.