NetApp hands storage operations to AI agents, but humans still draw the boundaries

NetApp wants AI agents inside the storage stack, but not running free. At NetApp INSIGHT, senior vice president and general manager of enterprise storage Sandeep Singh argued for a single data foundation that works across on-premises data centers, public clouds and neoclouds, instead of forcing customers into another silo.

His point is simple enough: enterprises are already spread across multiple environments, so the data infrastructure has to stay consistent even when the workload changes. Singh said customers need the same data capabilities and the same operating experience wherever their data sits. That consistency, he argued, is what makes production AI usable rather than just flashy.

Helen Yu, founder and chief executive officer of Tigon Advisory Corp., pushed the discussion toward management and accountability. She said the real problem is not storage, but leadership. When teams cannot trust the data, they work around official systems, and that opens the door to shadow AI and shadow IT.

NetApp’s answer is a unified control plane shared by humans and agents, delivered through NetApp Console. The idea is that customers set policies and boundaries, and agents keep infrastructure inside them. In one keynote demo, agents spotted a nighttime performance anomaly and applied quality-of-service rules without waking an engineer. Singh said the system kept the boundary intact, avoided a noisy-neighbor problem and left an audit trail for humans to review.

Yu’s take was more blunt: automation by itself is not governance. She said organizations need to fold machines and agents into their RACI charts, with clear answers for who owns the data, who can access it and who can override exceptions. That is the part enterprises keep skipping, and it is exactly where autonomous infrastructure either earns trust or becomes another clever headache.