Veeam Redesigns Data Resilience for the Era of AI Agents
Veeam Redesigns Data Resilience for the Era of AI Agents
When dozens of AI agents access data without going through a human interface for each employee, the boundary of security no longer coincides with the perimeter of the infrastructure and shifts to the data itself. It is around this idea that Veeam built the Milan stop of the VeeamON Tour 2026, the event during which they presented their latest developments to Italian partners and clients, just a few weeks after unveiling them at VeeamON New York. The event, hosted in a Milanese museum setting related to Leonardo and the history of science, took place in a symbolic year as the company celebrates its twentieth anniversary, having been founded in 2006 around the protection of virtual machines.
Twenty Years Later, the Third Era of Data
Welcoming the attendees was Alessio Di Benedetto, Country Manager for Italy at Veeam, who highlighted the scale the company has reached: over 550,000 customers in more than 150 countries, 80% of Fortune 500 companies, and a revenue exceeding $2 billion generated solely from software. This is the foundation from which Veeam is redefining its identity, transitioning from a backup specialist to a company focusing on data and AI trust, presenting itself as a Data and AI Trust Company. This strategic shift is backed by a concrete action: the $1.7 billion acquisition of Securiti AI, a specialist in data security management (Data Security Posture Management, or DSPM), completed at the end of 2025.
Explaining why this step was, in the company's words, necessary was Patrick Rohrbasser, Regional Vice President for South EMEA and Africa of Veeam. Following the backup and recovery era that began in the 1980s, and the ransomware defense era that emerged in 2017, a third era is now added: that of AI agents, changing the scale of the problem. On average, an organization coexists with 82 AI agents for each employee, and 97% of these operate with broader privileges than necessary; nearly nine out of ten companies already use them, but only 7% say they are fully capable of managing them. Complicating matters are the explosion of unstructured data, which today makes up 90% of the overall volume and is expected to triple in three years, and an attack speed that has outpaced human measures, with Microsoft blocking 7,000 password violation attempts per second, as noted by Rohrbasser.
The Missing Trust Level
The theme of the day is the gap between deploying AI and being able to trust it: according to Veeam, the former is now within everyone's reach, while the latter remains the real battleground.
"The infrastructure to deploy AI already exists. The one to ensure its trust does not," sums up the route of the group, stated CEO Anand Eswaran.
The company's response is the DataAI Command Platform, presented in New York and illustrated in Milan as the intermediary layer that connects infrastructure, data, and models. The platform brings together, in a single view, five areas that have so far been addressed in the market with different products, consoles, and vendors, namely security, governance, compliance, privacy, and data resilience. What holds them together is the DataAI Command Graph, the level of intelligence that maps where data resides, who accesses it, and how it is moved, a kind of social network of information that extends to both production data and backups.
The platform was previously announced in New York last May, and Edge9 reported on it in a previous article; in Milan, the discussion shifted to the rationale behind this change in approach. Focusing solely on agents, according to Veeam, would mean guarding what is known and losing sight of shadow AI, the tools developed outside IT control. Therefore, the point of observation remains the data, to be classified based on value and sensitivity, traced to who accesses it and for what purpose, and regulated at the source. The difference is illustrated by an example: a file with personal data in a shared but protected folder, accessible to a few known identities, is a manageable risk; the same file, the location or access of which is no longer known, becomes a toxic combination that, multiplied by the billions of documents present in a company, transforms into the real problem to tackle.
What Changes for Data Managers
On the product front, attention turned to the imminent Veeam Data Platform 13.1, which brings with it over 70 new features. The scope broadens to 11 platforms of on-premise virtualization, with the addition of Red Hat's OpenShift Virtualization alongside emerging alternatives to VMware. In terms of security, post-quantum encryption makes its debut, automated recovery of the entire Active Directory forest is introduced, and native support for identity management systems like Okta and Auth0, now among the most targeted.
However, the feature that best signals the direction is Intelligent ResOps, the first module built on the DataAI Command Platform, which brings together resilience and data management in a single console. For the first time, Veeam looks not only at backup data but also at production data: from a graphical map, one can select a virtual machine and see who accesses it, what geographical area it resides in, and which regulations apply to it, even restoring a single file with a natural language command. The entire logic is summarized in three verbs: detect, protect, and undo, aimed at a situation that is no longer theoretical—an agent accidentally deleting data. The first available integration will concern Microsoft 365.
The gap between adoption and control is not an abstraction, and in Europe, it weighs more than elsewhere. A study conducted by Veeam in the EMEA region reveals that while 99% of decision-makers consider data sovereignty critical, nearly three out of four organizations are sidelining it to accelerate AI adoption, and 40% identify the data used for AI and analytics as their primary blind spot. In a landscape regulated by GDPR, DORA, and the EU AI Act, being able to produce verifiable proof of compliance becomes an operational requirement, not just a statement of intent. It is on this ground that Veeam has reintroduced its Data and AI Trust Maturity Model, a free assessment path developed from discussions with over 300 CIOs and CISOs, which measures how prepared an organization is to govern AI and provides a benchmark with a map of priorities.
For a company that has built two decades of leadership in backup and recovery, shifting the focus to data and its security is a bet that reshapes the market and closely affects the ecosystem of Italian partners, called to guide clients in a territory where resilience, governance, and security are no longer separate disciplines. The message that emerged from Milan is that merely protecting data is no longer sufficient: it is necessary to demonstrate knowing where they are, who uses them, and how to restore them when something goes wrong, as this is the condition for adopting AI on a large scale without multiplying risk.