Aviz details Network Copilot deployment, connectors, and agent use

Aviz held APAC Network Copilot Office Hours focused on how teams can deploy Network Copilot, connect and normalize data, and use agents and LLM choices to support network operations. The session matters to enterprise IT and security leaders evaluating AI-assisted networking workflows and operational readiness.

Research Overview

The December (APAC) session was an online, interactive forum for asking questions and experimenting with Network Copilot. Aviz framed the format around live discussion rather than a slide deck or recorded content.

Presenters used an open Q&A structure and included a short, real-world demonstration of Network Copilot activities.

Key Findings

The discussion covered deployment and bringing Network Copilot up for on-site use, along with how roles, permissions, and projects can support running tasks in parallel. Attendees also asked about how data connectors operate, including attributes, normalization, MCP, and API approaches.

Aviz discussed how Network Copilot can use file uploads for context with RAG, and how agents can be used for operations. Questions also addressed available LLM options in Network Copilot and prompting techniques.

Technical Breakdown

The Q&A included specific implementation topics such as data connector mechanics and the role of normalization in connecting sources to the system. It also addressed MCP and API approaches as ways to integrate inputs.

On the agent side, the session covered agent options, the process for building custom agents, and an SDK walkthrough. The LLM portion covered choices available in NCP and the pros and cons that were discussed during the session.

Operational Impact

The included discussion described how traditional network operations toolsets can fragment information across dashboards, logs, alerts, and tools that do not share context. The session framed context switching as a factor that can reduce accuracy and increase manual correlation work.

Aviz described AI as augmenting engineers by synthesizing telemetry volumes, adding context to alerts, and helping teams reason through issues faster, while keeping engineers in decision-making roles. The Q&A also stated that clean, scalable data pipelines are a foundation for practical AI adoption in NetOps.

Blog Signals brief is a fact-based summary of the vendor blog.