‘What we’re focused on is this new era of vulnerability, risk, and exposure that AI causes, and that’s ushering in a new set of requirements for recovery. To manage AI recovery in light of some of these challenges, we have a whole set of innovations we’ll be announcing this week,’ says Tim Zonca, Commvault’s vice president of portfolio marketing.
Data protection and management technology developer Commvault Tuesday opened its annual Commvault Shift conference with the introduction of new technologies and strategic partnerships aimed at helping its channel partners modernize customers’ data infrastructure to meet the demands of AI.
This includes new ways to secure and recover AI agents; new ways to orchestrate the recovery of applications, identities, configurations, infrastructure dependencies, and datasets; and new partnerships with Nvidia, Zscaler, and Databricks.
The big theme of Commvault’s latest product introductions and partners, as well as the Shift conference as a whole, revolves around the fact that it is nearly impossible for even part of the day to avoid topics around AI, said Tim Zonca, vice president of portfolio marketing for the Tinton Falls, N.J.-based company.
“In particular, what we’re focused on is this new era of vulnerability, risk, and exposure that AI causes, and that’s ushering in a new set of requirements for recovery,” Zonca told CRN. “To manage AI recovery in light of some of these challenges, we have a whole set of innovations we’ll be announcing this week.”
Those challenges include:
- Data continues to proliferate across the enterprise and exist in more places across multiple data stores. “So the scope of what organizations need to understand, safeguard, protect, and make resilient is continuing to increase,” he said.
- Software is more and more likely to act on its own, making the management of AI agents more critical, as recent events have shown, Zonca said. That includes the possibility of agents either maliciously or unintentionally impacting businesses. “What hasn’t changed is there’s still mistakes and there’s still bad guys, but agents are acting independently,” he said.
- All these issues are now happening at machine speed, but traditional business processes were not built around that speed, he said.
Commvault Recovery Studio Brings Full-Stack Recovery Orchestration
To address those issues, Commvault Tuesday introduced a new set of capabilities called Recovery Studio, which Zonca said brings full-stack recovery orchestration to an organization. Recovery Studio is based on Commvault’s 2024 acquisition of Appranix, a provider of cloud cyber resilience technology, and embedded it in its Cloud Rewind platform.
“One of the really powerful things it does in the cloud world where everything is code and configuration was help move the industry beyond just recovering data,” he said. “You can now recover a full cloud application, even a full cloud or region or what have you, because it’s capturing not just the data, but also the application configurations, infrastructure configurations, network settings, security settings, metadata, and the underlying data.”
With Recovery Studio, Commvault is now applying that capability to traditional packaged applications, which Zonca said is really hard because they weren’t born in an era where everything is code.
“Recovery Studio allows you to sequence and automate and orchestrate a set of recovery actions that go just beyond the data,” he said. “It may include identity systems and access, the application constructs and configurations, the infrastructure configurations, the underlying data, and string this all together in a highly orchestrated manner. [The] result is extreme consistency, extreme reliability and repeatability, on a scale and a speed that then was previously not seen.”
This allows any workloads to be recovered in the same sort of way, and allows an organization to have automated runbooks to bring back kind of the full stack in the right order, Zonca said.
“It might be a cloud application, an on-prem application, and a set of services at a distribution center or colo, or something like that,” he said. “It might be a whole region. Regardless, it’s bringing back the entire stack that an organization needs to run there and deliver on their mission.”
Commvault’s push to use automation to protect and recover AI agents across a customer’s full IT stack and not just Commvault’s part via Recovery Studio is one of the ultimate goals for any customer, said Kevin Cronin, CEO of Kelyn Technologies, a Parker, Colo.-based solution provider and Commvault channel partner.
“The data is only one piece of it,” Cronin told CRN. “But the network is another piece, and the whole business operation or organizational operations is really the holy grail. Trying to figure it all out and put it all back together again is hard, like if Humpty Dumpty did fall off a wall, somebody’s got to put it all back together again.”
Commvault AgentRecover Expands AI Recovery To Microsoft Foundry, Amazon Bedrock AgentCore
Commvault also expanded its Commvault AgentRecover, the company’s full-stack AI recovery technology that helps businesses discover supported AI agents and dependencies, understand agent-driven changes, and recover components or connected AI systems to a trusted point.
AgentRecover looks for agents in a continual process, Zonca said. They might be Claude Code or OpenClaw, or perhaps agents built and deployed out of an AI platform like a Microsoft Foundry or Amazon Bedrock, he said.
“It looks for those agents and finds the tools that may be in place that these agents are using,” he said. “It maps dependencies around what tools the agents are using and what data sources are they touching. The underlying data might be Databricks or Snowflake or an Iceberg set of tables or Salesforce or some mix of these things. It maps the entire topology, and then if there’s some sort of rogue change, malicious or not, it understands the agent activity and can then roll everything back. Not just the agent or the affected agent systems, but everything, including all the dependencies.”
New to AgentRecover is support for Microsoft Foundry and Amazon Bedrock, which really round out its capabilities, Zonca said. Support for Google Vertex agents is on the roadmap, he said. There are also potential roadmaps to support external applications like ServiceNow, he said.
Cronin said his company partners with Amazon Bedrock, and he likes what he sees about AgentRecover.
“I think that goes a long way towards knowing who’s doing what in your environment and what they have access to,” he said. “This goes back to the question of, what’s the single source of truth for your data? I haven’t heard customers talk a lot about that yet. But it’s logical this will be something that they’re going to be concerned with. If they’re not yet, they will be.”
The whole AI landscape with regards to uptake is still for many people at the bleeding edge of IT, Cronin said.
“I think there’s a lot of people who are still dabbling, and starting to talk about MCP servers and things like that,” he said. “Having access to Commvault within your AI, or having Commvault have access to your infrastructure through AI, that’s a logical leap, but I don’t think most people have gotten there yet.”
Commvault Expands AI Partnerships With Nvidia, Zscaler And Databricks
A big theme of Commvault’s AI agent business is expanding its ability to work with strategic partners, Zonca said. These include:
Zscaler, with a focus on use cases such as AI discovery, runtime governance, and agent detection. Zscaler is a trusted provider with great prowess in identifying the users of AI applications and enforcing policy on what users and AI agents can do, and Commvault will work in concert with that intelligence to help recover the AI stacks, applications, and identities the agents use following an attack or outage, Zonca said.
Nvidia and Commvault are jointly developing a reference design to make protected enterprise data usable for AI while keeping it current, authorized, and traceable. The design builds on the data discovery and classification capabilities Commvault has added to its portfolio, which customers have typically used to understand and safeguard their data, Zonca said. Those same capabilities now let data scientists search for the data they need and flag sensitive content, such as PII (personally identifiable information), PHI (protected health information), and trade secrets, before it is shared with an AI data lake. Nvidia’s technology then handles translating the data and moving it to where AI systems can use it, he said.
Databricks and Commvault working on analytics. Commvault is deploying the same data activation capabilities with Databricks as it is with Nvidia, but here the goal is safely sharing data with Databricks, the underlying data store, for analytics and AI use cases, Zonca said. Commvault will also protect Azure Databricks as a workload, making it part of the full-stack recovery scenario when Databricks serves as the data layer for a business’s applications and services, he said.
Commvault Sees AI Recovery Opportunity For Channel Partners
Offerings built with strategic partners will, like all Commvault technologies, be offered through the company’s channel partners, said Alan Atkinson, the company’s chief business development officer.
Commvault is a channel-first company, and the new offerings are ready-made for channel partners, told CRN.
“In one case partners can couple a reference architecture, put services with it, and put it together,” he said. “In other cases, they can leverage their Databricks or Zscaler partnerships. Many of our channel partners are aggressively moving into the AI space and dealing with some of the gaps and barriers to wider adoption. For instance, what happens if an agent goes rogue? Being able to offer this along with the services we think accelerates partners’ ability to move more quickly into the AI space with all that’s associated with it.”
