Rubrik (NYSE:RBRK) has introduced Rubrik Agent Cloud for Anthropic’s (NASDAQ:ANTP) Claude Code and Claude Cowork, providing organizations with new capabilities to deploy, monitor and manage AI-powered agents while maintaining greater operational control.
The launch includes tools designed to track agent activity, recover from unintended actions and strengthen resilience as enterprises increasingly adopt autonomous AI systems.
New Safeguards for Claude-Powered Agents
The platform introduces what Rubrik describes as an agent rewind capability, allowing organizations to reverse erroneous actions performed by AI agents.
The offering also includes immutable codebase recovery, enabling businesses to restore systems when mistakes bypass traditional version-control protections.
Rubrik said the platform continuously backs up, monitors and restores the configurations that govern agent behavior, helping organizations maintain oversight as AI agents take on more complex operational responsibilities.
Autonomous Business Recovery Introduced
Alongside Agent Cloud, Rubrik announced Autonomous Business Recovery for Cloud Applications, a solution designed to restore cloud environments that include data, networking, identity systems and application configurations.
The technology is intended to accelerate recovery efforts by rebuilding what the company refers to as an organization’s Minimum Viable Business through highly automated processes.
Rather than focusing on individual systems, the platform aims to restore the interconnected infrastructure required to resume critical business operations following a disruption.
Addressing Emerging AI Security Risks
Rubrik noted that AI agents are increasingly capable of writing, committing and deploying code without direct human supervision.
However, many enterprise security frameworks were originally designed with the assumption that people would remain involved in key decision-making processes.
According to the company, this shift introduces new risks including rogue code commits, repository ransomware, prompt injection attacks and intellectual property theft, all of which can occur at speeds beyond the capabilities of many traditional DevSecOps controls.
“Organizations are adopting Claude faster than any agentic technology we have seen, and every security leader asks the same question: how do we stay in control when an agent can act?” said Anneka Gupta, Chief Product Officer at Rubrik.
Recovery Planning Conducted Before Attacks Occur
The Autonomous Business Recovery platform is powered by Rubrik’s Preemptive Recovery Engine, which performs recovery preparation activities during normal business operations.
The system identifies critical applications and their dependencies, validates clean recovery points and builds recovery plans in advance.
As a result, when a cyberattack or operational failure occurs, recovery workflows are already prepared, potentially reducing downtime and accelerating restoration efforts.
Growing Concern Over Recovery Readiness
Research from Rubrik Zero Labs indicates that 88% of business leaders are concerned about their ability to meet existing recovery time objectives as AI-driven threats become more sophisticated.
The company argues that traditional backup solutions often focus on restoring individual assets rather than the complex web of dependencies that underpin modern cloud applications.
With organizations increasingly deploying autonomous AI systems, Rubrik believes that more comprehensive recovery and governance capabilities will become essential components of enterprise cybersecurity strategies.