NETSCOUT has added Model Context Protocol (MCP) connectivity to its Omnis AI Insights platform, allowing AI assistants and agents to access curated network data directly rather than relying solely on information passed through conventional monitoring and analytics systems.
The company said the capability is intended to give AI systems more reliable operational context when investigating network and application performance, security incidents and other IT issues.
MCP, an open protocol increasingly being adopted as a way for AI models and agents to interact with external tools and data sources, provides the connection between NETSCOUT’s data platform and AI assistants or agents.
The new capability sits within Omnis AI Insights and uses NETSCOUT’s Adaptive Service Intelligence (ASI) technology to extract and contextualise network information at the point where it is collected. The resulting ‘AI-ready Smart Data’ is designed to retain operational context while being more compact than raw network telemetry.
NETSCOUT said its Omnis Sensor performs semantic extraction at network vantage points, capturing information about applications, services, transactions and user behaviour. Omnis Streamer then collects and curates the resulting data, with configurable playbooks allowing datasets to be tailored to different operational requirements.
The Streamer can now expose this data through a built-in MCP server, allowing AI assistants and agents to request relevant network evidence at run time. NETSCOUT also supports integrations with platforms including Splunk, the ELK Stack, Datadog, ServiceNow and Dynatrace.
The company argues that processing and contextualising network data before it reaches an AI model can reduce the volume of telemetry that needs to be processed, potentially lowering token and infrastructure costs while giving AI systems more specific evidence to work with.
Phil Gray, AVP of Product Management at NETSCOUT, said the addition of MCP complements the company’s existing Kafka streaming capabilities by allowing the same data to be accessed directly by AI models and agents.
NETSCOUT cited a deployment in which conventional application monitoring tools detected no application errors even as network conditions were affecting user experience. Its Smart Data captured network-level evidence including minimum TCP window size, retransmission counts and zero-window events, which the company said allowed an AI system to distinguish observed conditions from assumptions.
There’s plenty of other editorial on our sister site, Electronic Specifier! Or you can always join in the conversation by commenting below or visiting our LinkedIn page.
