Commvault announces trio of AI tools for data prep, agent oversight, resilience

Commvault has introduced three AI-focused offerings: one to prepare backup data for AI use, one to discover and monitor AI agents, and one to build cyber-resilience agents.

It recognizes that large language models and agents need access to safe, reliable data, and that AI agents, like people operating within an IT estate, need their privileges and actions overseen and rolled back if something goes wrong. It also understands that Commvault users will want to use their own cyber-resilience and recovery agents. The Data Activate, Agent Protect, and AI Studio offerings aim to provide these three functions,  with Commvault Cloud as the environment in which they operate.

Commvault president and CEO Sanjay Mirchandani said: "Every enterprise era has produced a system of record – ERP for business operations, CRM for customers, and now AI for the enterprise. 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." 

Data Activate, previously known as Data Rooms, enables organizations to classify and curate data from protected backup copies, across on-premises and cloud environments, and prepare approved datasets in formats like Apache Iceberg and Parquet for use with large language models and AI data platforms. It integrates with Microsoft Azure and Snowflake through support for Parquet and Iceberg. Staff can identify and exclude personally identifiable information before activating datasets for model development.

It creates governed, policy-controlled "rooms" inside Commvault Cloud – spaces where data can be classified, curated, and shared with AI and analytics tools without leaving the protection boundary. Data Activate can regularly publish updated datasets, and is, in effect, an ETL workflow to feed AI pipelines with the most current trusted data.

AI Protect

AI Protect will discover and inventory agents across environments and map their activity to AI stacks. It will be designed to discover AI agents (and their dependencies) operating across connected environments on a recurring basis, helping maintain a unified, up-to-date inventory based on configurable discovery cadence. Each agent record will capture its execution environment and the data sources, models, configurations, applications, and infrastructure with which it interacts. It will help provide a complete, cross-environment picture of what’s running and what it touches.

Agent activity can generate weak risk signals. Agents can generate enormous telemetry, but without correlation across identity, access, and impact, distinguishing benign automation from high-risk behavior remains a manual effort.

AI Protect will be designed to surface protection status for every agent-accessed asset – protected, partially protected, or not protected – and help identify gaps introduced by agent activity. Where gaps exist, it will offer recommended actions and protection workflows to enable teams to close them.

It will ingest agent activity from existing audit, event, and telemetry sources, and present it in agent-centric context – not as raw logs. A time-ordered activity timeline will show what each agent has done and when, and risk signals will be automatically flagged and categorized when agents access sensitive data, interact with unprotected assets, or exhibit unusual patterns.

When an agent-initiated change causes an issue, AI Protect will surface recovery point availability for impacted assets and guide teams through the appropriate recovery action – whether that's restoring data, applications, or configurations. Recovery will be scoped directly to the agent's impact, not generic incidents, and every action will be time-stamped.

In addition, teams will be enabled to recover the full AI stack – not just the model, but the connected data, configurations, and underlying systems that support it – helping restore the entire environment to a known good state with a single, guided action.

Commvault CTO Pranay Ahlawat said: "In agentic environments, agents mutate state across data, systems, and configurations in ways that compound fast and are hard to trace. When something goes wrong, teams need to recover not just data, but the full stack – applications, agent configurations, and dependencies — back to a known good state. That's what AI Protect delivers."

AI Studio

AI Studio will enable customers to create and utilize agents that address their specific needs, including a repository of built-in agents for common resilience use cases, such as backup, recovery, and incident response workflows. Teams will be able to build custom agents that automatically and securely use Commvault's Model Context Protocol (MCP) server and integrate with other enterprise systems.

AI Studio has three components: Agent Library, Agent Builder, and Agent Management.

The Agent Library will be the entry point to AI Studio. It will bring AI-enabled automation to data protection and recovery. Through Arlie, users can interact with specialized agents that analyze, recommend, and act – turning complex resilience workflows into simple, conversational experiences.

It will present a structured inventory of every agent available in the environment – both default agents built by Commvault and custom agents created by the customer – grouped by type and showing each agent's name, category, description, and enabled status at a glance. The Agent Library will offer teams a single, authoritative view of their resilience agent ecosystem before taking any action.

Its agents work through Arlie's intuitive interface, helping teams simplify resilience management and accelerate outcomes through conversational, intelligent automation. Default agents include:

• Arlie Data Sense
• Arlie Advisor
• Arlie Classify
• Arlie Insights
• Arlie Recover 

AI Studio's Agent Builder will enable administrators to create custom agents by leveraging Commvault's workflows and MCP server. An administrator will be able to describe what they want to automate – for example: "I need an agent that detects when storage or infrastructure issues are starting to impact backups and helps resolve them before they affect SLAs."

The system will be designed to translate that intent into a structured agent configuration, including triggers, conditions, and actions, with optional AI-enabled steps from Arlie – such as Summarize, Generate Recommendation, or Draft Notification – available as explicit workflow steps.

The administrator will be able to review the proposed workflow, adjust it as needed – changing trigger frequency, specifying a distribution list, or reordering steps – and save it. The result will be an auditable custom agent that appears in the Agent Library and can be managed through Agent Management like any other agent.

AI Protect and AI Studio are future offerings. To learn more about Commvault's latest AI offerings, read the company's blogs about Data Activate, AI Protect, and AI Studio.