Network and Infrastructure: AI Needs a Semantic Map – Netbox Labs Triple-T
LLMs and agentic patterns have upleveled AI operations beyond traditional big data and ML approaches, layering on reasoning and tool use to connect AI directly with the operations stacks of infrastructure teams. But “throwing LLMs at the problem” isn’t a successful approach.
Agents need context to make sense of metrics, logs, configs, and other operational data. They need a “semantic map”- a structured data model, accessible via agentic tool use, empowers LLMs with information about the topology and components of the infrastructure. A semantic map enables agents to make sense of observability and operations data, plan and execute effective investigations, and effectively share information with human operators and other agents.
In this talk, we’ll break down how we leveraged NetBox, the most widely deployed data model for documenting infrastructure, as the semantic map for modern AI operations. We’ll share lessons from our work with open source and commercial AI products, break down real-world agentic patterns, and demonstrate how semantic maps enable faster, more effective, and genuinely useful AI operations.
Kris Beevers is the Co-founder and CEO of NetBox Labs. Previously, Kris co-founded and led NS1, which was acquired by IBM in 2023. He has founded and helped lead multiple other successful businesses in internet and network infrastructure. Kris is an engineer at heart and by background, and loves the leverage infrastructure innovation creates to accelerate technology and empower engineers to do their best work. He holds a Ph.D. in Computer Science from RPI and is perennially interested in too many things.
