Rubrik announced a series of AI-focused security products during its FORWARD user conference. At the heart of this is Rubrik AI, which transforms the company’s platform into an autonomous AI agent that performs security tasks independently.
In addition, Rubrik is introducing Agent Cloud for Claude Code and Claude Cowork from Anthropic, designed to give organizations more control over AI agents that develop software.
According to Rubrik, the rise of autonomous AI agents is creating new security challenges. While existing security solutions rely on human intervention, AI agents can write, modify, and deploy code on their own. As a result, errors or attacks can spread much more quickly.
With Rubrik AI, the vendor aims to further automate security processes. Organizations specify the desired outcome, after which the platform independently determines the necessary steps. In doing so, the software takes into account the IT environment and current security risks.
CEO and co-founder Bipul Sinha presents the new approach as a response to both attacks that leverage AI and the risks arising from organizations’ increasing use of AI agents themselves. According to him, cyber resilience must keep pace with the speed at which these systems operate.
Rubrik AI consists of several components. Agentic Mode uses a single AI agent for Rubrik Security Cloud and Rubrik Agent Cloud and analyzes data, identities, and deployed AI agents in conjunction. Guardrails ensure that every autonomous action remains verifiable and can be reversed if necessary. In addition, the platform automates recovery procedures that were previously performed largely manually.
Security for Claude Code
In addition to Rubrik AI, the company is introducing Rubrik Agent Cloud (RAC) for Claude Code and Claude Cowork. This expansion is aimed at development teams that use AI agents for software development.
One of the most notable features is Agent Rewind. This allows organizations to undo unintended actions by AI agents, even when standard version control systems do not provide a solution. Rubrik cites examples such as overwriting the commit history or deleting all branches in a repository.
To achieve this, RAC continuously stores immutable snapshots of GitHub and Azure DevOps repositories outside the development environment. If an AI agent or attacker causes damage, a repository or an entire development organization can be restored to a previous recovery point.
In addition, Rubrik is introducing the Semantic AI Governance Engine (SAGE), which manages AI agents in real time based on policy rules. A centralized inventory provides insight into access rights, risks, and any policy violations.
The configuration of Claude agents is also protected. Rubrik backs up system prompts, tool permissions, and configuration files, continuously monitors for unauthorized changes, and can automatically revert to a previous configuration in case of deviations.
With these announcements, Rubrik is targeting organizations that not only use AI agents as assistants but also allow them to perform operational tasks independently. According to the company, this development requires a security layer that protects not only data but also the AI agents themselves and the actions they perform.
Also read: ‘Fix this code’: three words behind the export ban on Claude Fable 5