Rubrik introduced a sequence of AI-focused safety merchandise throughout its FORWARD consumer convention. On the coronary heart of that is Rubrik AI, which transforms the corporate’s platform into an autonomous AI agent that performs safety duties independently.
As well as, Rubrik is introducing Agent Cloud for Claude Code and Claude Cowork from Anthropic, designed to provide organizations extra management over AI brokers that develop software program.
In line with Rubrik, the rise of autonomous AI brokers is creating new safety challenges. Whereas present safety options depend on human intervention, AI brokers can write, modify, and deploy code on their very own. Consequently, errors or assaults can unfold way more shortly.
With Rubrik AI, the seller goals to additional automate safety processes. Organizations specify the specified end result, after which the platform independently determines the required steps. In doing so, the software program takes into consideration the IT surroundings and present safety dangers.
CEO and co-founder Bipul Sinha presents the brand new strategy as a response to each assaults that leverage AI and the dangers arising from organizations’ growing use of AI brokers themselves. In line with him, cyber resilience should preserve tempo with the velocity at which these programs function.
Rubrik AI consists of a number of elements. Agentic Mode makes use of a single AI agent for Rubrik Safety Cloud and Rubrik Agent Cloud and analyzes information, identities, and deployed AI brokers in conjunction. Guardrails make sure that each autonomous motion stays verifiable and will be reversed if mandatory. As well as, the platform automates restoration procedures that had been beforehand carried out largely manually.
Safety for Claude Code
Along with Rubrik AI, the corporate is introducing Rubrik Agent Cloud (RAC) for Claude Code and Claude Cowork. This growth is aimed toward improvement groups that use AI brokers for software program improvement.
One of the notable options is Agent Rewind. This enables organizations to undo unintended actions by AI brokers, even when normal model management programs don’t present an answer. Rubrik cites examples similar to overwriting the commit historical past or deleting all branches in a repository.
To realize this, RAC constantly shops immutable snapshots of GitHub and Azure DevOps repositories outdoors the event surroundings. If an AI agent or attacker causes harm, a repository or a whole improvement group will be restored to a earlier restoration level.
As well as, Rubrik is introducing the Semantic AI Governance Engine (SAGE), which manages AI brokers in actual time primarily based on coverage guidelines. A centralized stock supplies perception into entry rights, dangers, and any coverage violations.
The configuration of Claude brokers can also be protected. Rubrik backs up system prompts, instrument permissions, and configuration information, constantly screens for unauthorized adjustments, and might routinely revert to a earlier configuration in case of deviations.
With these bulletins, Rubrik is concentrating on organizations that not solely use AI brokers as assistants but additionally permit them to carry out operational duties independently. In line with the corporate, this improvement requires a safety layer that protects not solely information but additionally the AI brokers themselves and the actions they carry out.
Additionally learn: ‘Repair this code’: three phrases behind the export ban on Claude Fable 5