Rubrik unveiled three new offerings at its global Forward events: Rubrik AI, Rubrik Agent Cloud (RAC) built for Anthropic’s Claude Code and Claude Cowork, and Rubrik Autonomous Business Recovery for Cloud Applications. The vendor stressed AI agent threats must be mitigated at machine speed, as manual human response can never keep pace with fast-evolving automated attacks.
Rubrik AI works as a unified intelligent agent across Rubrik Security Cloud (RSC) and Rubrik Agent Cloud (RAC). Powered by technology from its Predibase acquisition, RAC serves as a dedicated operation and governance suite. RSC is Rubrik’s core SaaS platform covering data security, resilience, observability and end-to-end recovery. Meanwhile, Rubrik Autonomous Business Recovery for Cloud Applications (RABRCA) delivers full protection and one-stop recovery for complete cloud business services and integrated application stacks.
Varun Grover, Director of Product Marketing for AI and SaaS at Rubrik, emphasized Rubrik AI is not a simple chatbot add-on, but a brand-new interactive paradigm for Rubrik’s platform. “Enterprises only need to define expected business outcomes — cleaning compromised environments, narrowing attack blast radius and restoring normal business operations. Rubrik AI will analyze on-site infrastructure and complete remediation actions autonomously to meet targeted goals,” he said.
For cloud recovery scenarios, Rubrik AI automatically locates the last healthy data state, generates tailored recovery plans based on application dependency mapping, and restores compute, network and data resources in correct order without manual human approval at each step. It cuts recovery time from weeks via manual operation down to mere minutes.
Grover added that Rubrik AI’s Agentic Mode enables cross-environment analysis covering core data, user identity and all deployed enterprise agents. Instead of treating these tools as isolated siloed systems, the platform monitors all agent behaviors, analyzes internal correlations and takes coordinated response actions when emergencies occur. Built-in Agentic Guardrails guarantee all autonomous operations are fully auditable, traceable and reversible.
RAC for Claude Code & Claude Cowork
The RAC solution for Anthropic’s Claude tools enables large-scale deployment of Claude-based AI agents, with full-stack observability, unified governance and one-click agent action rewind. It can restore immutable codebases even when version control fails to roll back harmful operations, building a solid resilience layer for generative AI agents.
Rubrik pointed out that autonomous AI agents can independently write, submit and deploy code nowadays. However, they may also trigger critical security risks including malicious code commits, source code ransomware attacks, prompt injection and intellectual property theft at machine speed, whose damage scope far exceeds the defense capability of traditional security tools.
Anneka Gupta, Chief Product Officer at Rubrik, commented: “Enterprises are adopting Claude agents faster than any previous agentic tool. All security leaders face the same core concern: how to retain full control over autonomous AI agents. Rubrik Agent Cloud adds a critical resilience layer for Claude, allowing teams to monitor agent access permissions, regulate agent behaviors, reverse harmful operations, and restore corrupted code and agent configuration files.”
Core Control Capabilities of Rubrik’s Claude Environment
-
Semantic AI Governance Engine (SAGE): Real-time security and governance for autonomous agents, replacing rigid manual rules with intent-based management to safely expand enterprise AI deployment scale.
-
Agent Inventory: Provides full visibility of risk exposure, access privileges and policy violations across all deployed AI agents.
-
Agent Rewind: Instantly roll back unintended behaviors from all mainstream agents, including customized agents and tools built on Claude Code and Claude Cowork.
-
Codebase Resilience: Creates immutable snapshots for GitHub and Azure DevOps repositories stored outside original code libraries. It supports one-click recovery for irreversible damages such as full branch deletion and forced commit overwrites, including code ransomware rollback.
-
Resilience for Claude Agents: Backs up and tracks version changes of core Claude agent configurations such as system prompts, tool permissions and key setup files across organizational, repository and user levels. It monitors abnormal configuration drift in real time and links faulty changes to healthy backup snapshots for fast autonomous recovery.
Click the link to learn more about RAC for Claude Code and Cowork.
Rubrik Autonomous Business Recovery for Cloud Applications (RABRCA)
RABRCA functions as an autonomous decision engine, transferring complicated attack forensics and coordinated recovery workflows from human security teams to intelligent software. Its key features are as follows:
-
Autonomous Application Discovery: Generates dynamic graphical application dependency maps. Users only need to tag one core anchor resource such as an EC2 instance, and the platform recursively maps the whole infrastructure with real-time updates, eliminating outdated static architecture diagrams during security incidents.
-
Synchronized Immutable Protection: Delivers gap-free continuous protection covering compute, network and data layers. Newly added cloud resources are automatically included in air-gapped, immutable SLA protection policies to avoid blind protection gaps.
-
Orchestrated Clean Application Recovery: Its Preemptive Recovery Engine pinpoints healthy system states in seconds without temporary auxiliary infrastructure. It executes validated recovery plans following dependency sequences, and supports non-disruptive pre-recovery testing and compliance report generation.
Users can access the official RABRCA solution brief via the attached link. RABRCA is currently in private preview, and Rubrik is recruiting design partners for joint testing. Rubrik AI has been officially released for global customers.
Industry Comment
We believe mainstream data protection and cyber resilience vendors will universally adopt agent-based automated protection and recovery management, while adding dedicated defense and resilience functions for third-party AI agents such as Anthropic Claude. For vendors including Cohesity, Commvault, Druva, HYCU and Rubrik, such agent-native security capabilities will become essential industry standard configurations.
Bootnote: Safe Harbor Statement
Unreleased services and features mentioned in this material are not currently available for commercial use. Rubrik reserves full discretion to adjust release schedules or cancel pending functions without prior notice. All forward-looking descriptions do not constitute official delivery commitments, contractual obligations or binding promises. Customers shall make purchase decisions solely based on currently available general-release products and features.
Beijing Qianxing Jietong Technology Co., Ltd.
Sandy Yang/Global Strategy Director
WhatsApp / WeChat: +86 13426366826
Email: yangyd@qianxingdata.com
Website: www.qianxingdata.com/www.storagesserver.com
Business Focus:
ICT Product Distribution/System Integration & Services/Infrastructure Solutions
With 20+ years of IT distribution experience, we partner with leading global brands to deliver reliable products and professional services.
“Using Technology to Build an Intelligent World”Your Trusted ICT Product Service Provider!
Your message must be between 20-3,000 characters!
English
