Veeam introduces Agent Commander, a solution for detecting AI risks, protecting AI systems, and undoing AI errors. The tool is the result of the acquisition of Securiti AI at the end of 2025.
Agent Commander is designed to help organizations proactively address AI-driven risks while securely scaling AI agents. “AI happens at machine speed, which means organizations must understand what data is being used, by what agent, and how in real-time,” said Anand Eswaran, CEO of Veeam. “If an error occurs, organizations not only need to understand what data was impacted, but they also need the ability to undo any damage rapidly.”
Trust as the biggest challenge
The most critical gap in today’s AI infrastructure, according to Veeam, is trust. As AI agents scale, data risk and AI risk have become one and the same. An agent is only as trustworthy as the data it can see, access, and act on. Yet enterprise controls remain fragmented with separate systems for protection, security, governance, and recovery.
Sensitive data is fed into models and acted upon in ways that no one has approved or is tracking. Because AI moves very quickly, an AI agent can access and act on sensitive data in seconds. Traditional workflows that take hours to detect and days to recover leave too much exposure at that speed.
Data Command Graph at the core
At the heart of Agent Commander is Veeam’s Data Command Graph, a real-time relational intelligence engine that maps live connections between data, identities, AI models, and autonomous agents. This happens in both production and backup environments.
Veeam sees the toxic combinations where compromised identities, exposed data, and autonomous agents intersect. The platform also shows how those risks worsen, spread, and evolve in real time across the entire data and AI landscape.
Agent Commander offers three features that make it unique. First, it detects AI risk in context by identifying shadow AI, sensitive data exposure, and risky agent behavior. Second, it autonomously protects AI pipelines by enforcing detailed, real-time controls over data, identities, and AI agents. Third, it can precisely undo AI errors by surgically reversing unwanted AI actions.
Agent Commander unifies control over production and backup for the detection and enforcement of granular policies. Organizations can request early access via the Veeam website. Agent Commander will be available in a future release of the Securiti Data Command Center.