Rubrik has launched Rubrik AI, introducing an agent-first interface across its platform.
The new offering spans Rubrik Security Cloud and Rubrik Agent Cloud. Users can set a business outcome, and the software will carry out tasks aimed at achieving it.
The launch marks a shift in how customers interact with Rubrik's software portfolio. Rather than navigating individual tools directly, users are meant to work through a single agent operating across data, identity and other areas of the platform.
The system is designed to adapt to an organisation's context and security threats, and to act autonomously across Rubrik's products. Rubrik presented the approach as part of a broader effort to address cyber incidents and AI-related risks that unfold at machine speed.
Features outlined by Rubrik include Agentic Mode, which uses one agent across Rubrik Security Cloud and Rubrik Agent Cloud, and Agentic Guardrails, built-in controls intended to make autonomous actions auditable, attributable and reversible.
Rubrik also highlighted Orchestrated Workflows, which can automate multi-step recovery processes that would otherwise require extensive manual work by security and IT teams.
Platform shift
The launch adds another layer to Rubrik's push beyond backup and recovery software into a broader mix of cyber resilience, identity protection and AI oversight. The company has increasingly framed its products around the need to protect not only stored data and workloads, but also the software agents organisations are beginning to deploy in day-to-day operations.
The announcement fits within a wider trend in enterprise technology, where software groups are introducing AI agents that can interpret a user's request, make decisions and trigger actions across systems with less human input. At the same time, these moves have drawn fresh attention to governance, audit trails and the risk of unintended automated actions.
Rubrik said its controls are intended to prevent what it described as runaway AI risks. It linked this to the need for organisations to manage both external attacks and risks tied to internal AI agent deployments.
Bipul Sinha, Chief Executive Officer, Chairman and Co-Founder of Rubrik, described the launch as a change in the role of the company's software.
"Today, Rubrik becomes an agent. Our Agentic Cyber Resilience is designed to mitigate risks from both external AI attacks and internal agent deployments," said Bipul Sinha, Chief Executive Officer, Chairman and Co-Founder, Rubrik.
He also outlined the company's view of how the product will be used in cyber response and recovery.
"Rubrik AI will deliver agentic automation to enable resilience against machine-speed cyber breaches and compromised AI agents," said Sinha.
Broader push
The product launch was announced alongside a series of other updates across Rubrik's portfolio, including work related to unstructured data, cloud application recovery, identity resilience and support for Anthropic's Claude Code through Rubrik Agent Cloud.
Those announcements suggest Rubrik is trying to position itself around a broader operating model in which AI systems are not only tools to assist analysts, but active participants in security, recovery and governance processes.
For customers, the practical question will be how far they are prepared to let software take autonomous action in environments that include sensitive data, business-critical workloads and identity systems. Rubrik's emphasis on reversibility and auditability suggests that trust and oversight remain central to adoption, especially where automated responses may affect live operations.
Rubrik said its design for Rubrik AI is patent pending and that the software is built to reason, act and recover at machine speed.