NETSCOUT expands data platform to boost enterprise AI

NETSCOUT has expanded its data platform for enterprise AI, aiming to improve the operational context available to AI systems.

The announcement centres on a platform that observes digital interactions and converts network packets into what NETSCOUT calls contextualised operational evidence in real time. The information is then curated for uses including observability, service assurance, cybersecurity and AI-driven operations.

NETSCOUT argues that many enterprise AI projects struggle because the data supplied to models is incomplete or lacks context. In network and IT operations, organisations often rely on metrics, events, logs and traces, known as MELT data. That can force AI systems to reconstruct incidents after information has already been sampled, aggregated or split across different tools.

According to the company, this increases inference and compute demands, as well as token consumption and the risk of inaccurate recommendations. NETSCOUT is positioning its platform as an alternative source of operational evidence drawn directly from network traffic.

Data focus

At the core of the approach is deep packet inspection, used to derive meaning from packets at the point of observation. NETSCOUT says this preserves evidence that can be lost in more conventional datasets and gives AI systems denser, more relevant context from the start.

The platform is designed to sit alongside existing observability tools rather than replace them. NETSCOUT says the data can also be integrated into broader enterprise data and AI workflows, including analytics platforms, large language models, copilots and AI agents.

This reflects a broader market shift as companies look beyond model choice and focus more on the quality of the data fed into those systems. Industry analysts have increasingly argued that AI tools need access to real-time, contextual and broadly accessible data to support operational decision-making reliably.

To support its case, NETSCOUT cited internal testing. It reported more than a 25% reduction in AI token consumption compared with using MELT-only data, alongside more than a 75% reduction in mean time to know, a measure of how quickly teams can determine what happened in an incident.

Operational use

The platform is aimed at several operational groups, including network operations, security operations, development operations, site reliability engineering and service teams. NETSCOUT says these users could gain natural-language access to detailed operational evidence to speed up investigations and problem resolution.

Another part of the pitch is cost control. By increasing the density of relevant information, NETSCOUT says customers can reduce the volume of low-value telemetry that AI systems must process, cutting storage, token and compute costs without losing the context needed to understand service behaviour.

The company is also framing the platform as a step towards more automated IT operations. It says independently observed and explainable evidence can help support governance and audit requirements as organisations move from AI-assisted recommendations to more autonomous actions.

This matters because many large organisations are experimenting with AI agents in infrastructure, security and service management, but remain cautious about letting those systems act without human oversight. In that context, the quality and traceability of underlying operational data becomes a key issue.

NETSCOUT says its common data foundation can support use cases across hybrid, multi-cloud, containerised, virtual and physical environments. It says the platform can help teams identify hidden dependencies, separate infrastructure failures from application problems, detect protocol and security exposures, and understand the operational effect of an event across services.

Market position

The announcement also shows how NETSCOUT is trying to extend its established network visibility technology into the growing market for AI operations and automation. As access to models broadens and technical differences between them narrow, suppliers across the sector are increasingly competing on data quality, context and integration with existing operational workflows.

That could appeal to customers that want to add AI to current systems without a full overhaul of their tooling. It may also interest technology partners looking for network-derived operational data to feed into analytics and automation systems.

Sanjay Munshi, Chief Operating Officer, NETSCOUT, outlined the company's case for the approach. "Unlocking the benefits of AI across the enterprise will not be achieved by adding another model. It will succeed through context engineering: giving AI the right operational context before reasoning begins," he said.

He also pointed to the company's internal testing and its view of where the platform fits in IT operations. "NETSCOUT turns observed digital interactions into grounded-truth evidence. Through our own internal testing we experienced more than a 25% reduction in AI token consumption compared with MELT only data, and more than a 75% reduction in MTTK. Compact, context-rich operational intelligence helps our customers improve decision confidence, lower the cost of AI-driven analysis, and establish the control required to move from AIOps recommendations toward safe, autonomous operations," Munshi said.