Data protection provider Commvault Systems Inc. today announced new artificial intelligence capabilities that it says will allow organizations to accelerate AI adoption with confidence while maintaining control over data, agents and recovery.

The new capabilities allow enterprises to activate AI safely, discover and govern AI agents and build and control agentic workflows from within Commvault Cloud.

As organizations deploy AI, many are struggling to balance innovation with the risks of exposing sensitive data and losing control over AI agents. According to data from Deloitte Touche Tohmatsu Ltd. cited by Commvault, 60% of AI leaders cite risk and compliance concerns and legacy system integration as the top barriers to agentic AI adoption.

The new offerings — Data Activate, AI Protect and AI Studio — address the challenges faced by organizations deploying AI by helping them understand the impact of agent-driven changes, roll them back when necessary and de-risk AI.

“Every enterprise era has produced a system of record — enterprise resource planning for business operations, customer relationship management for customers and now AI for the enterprise,” said Chief Executive Sanjay Mirchandani. “If data powering AI is compromised, AI is compromised. If data can’t be recovered, AI can’t be trusted. Commvault Cloud is the system of record for AI resilience.”

The first new release, Data Activate, allows organizations to classify and curate data from protected backup copies and prepare approved datasets in formats like Apache Iceberg and Parquet for use with large language models and AI data platforms. It continuously publishes updated and vetted datasets, making it easier to keep AI pipelines in sync with trusted data.

AI Protect, meanwhile, helps organizations identify risk, understand the impact of agent-driven changes, recover affected applications and perform full-stack recovery across AI-driven environments. It discovers and inventories agents across environments, maps their activity to the AI stacks they interact with, such as vector databases and surfaces high-risk behavior. When issues occur, teams can use the AI Protect to assess the impact quickly and initiate guided, one-click, complete recovery workflows to restore affected data, systems and configurations to known good states.

Finally, AI Studio enables organizations to create and use agents that address their specific needs. It includes a repository of built-in agents for common resilience use cases and allows teams to build custom agents that automatically and securely use Commvault’s Model Context Protocol server and integrate with other enterprise systems.

Combined, Commvault says, the three AI capabilities allow teams to advance an AI resiliency lifecycle by safely activating data for AI purposes all the way through bringing discovery, governance, protection and recovery to AI agents across on-premises, software-as-a-service and hybrid cloud environments.

Mirchandani spoke with theCUBE, SiliconANGLE Media’s livestream studio, in February, when he discussed how Commvault is reframing cyber resilience at a time when AI-driven systems accelerate the pace of change across data, applications and access controls.

Photo: Commvault