# Aviz Networks outlines Network Copilot 2.0 release

URL: https://decisioninsights.ai/aviz-networks-outlines-network-copilot-2.0-release/

Aviz Networks has released Network Copilot 2.0, an agentic AI platform for operating network and GPU infrastructure through natural language. The update adds GPU, inference, and cluster integrations and voice support in 18 languages for enterprise operations teams.Product OverviewNetwork Copilot is described as an autonomous agentic AI platform for network and AI infrastructure operations. It connects to existing network devices, controllers, monitoring, security, ITSM, and GPU systems through a common natural language interface without requiring a new data store or data migration. Users can build and embed their own workflows with the Agent SDK.Release 2.0 CapabilitiesR2.0 adds NVIDIA Dynamo, DCGM, and BCM integrations, Prometheus-based metrics, troubleshooting agents, and conversational fabric and tenant operations. Operators can interact through text or voice in 18 languages. Results can be presented as conversational answers, heatmaps, network graphs, charts, and dashboards.Automation and GovernanceNetwork Copilot starts with interactive troubleshooting and purpose-built queries, and background agents can automate recurring operations while requiring human confirmation for actions that need operator approval. Custom agents are built and tested in a sandbox using the Agent SDK and share the reasoning engine and connectors used by built-in agents. Agent Trace shows the agents, tools, and sources behind each response, while searchable audit logs and RBAC controls support governance.Security and Compliance AuditsThe platform runs 59 evidence-based security checks with findings mapped to NIST SP 800-53 controls. Users can create custom audit templates, compare results against previous audits, and export findings to CSV or Excel. Activity logs can be exported as CSV, JSON, or NDJSON and forwarded to a SIEM.Deployment and Commercial TermsNetwork Copilot can be deployed on-premises with an air-gapped private LLM, in a private cloud, or through a GPU-as-a-Service quick start, and it can connect to supported public LLM services where data policies permit. It is offered as an annual software subscription with 1-, 3-, and 5-year terms. Aviz-developed and supported agents are available as add-on options.Aviz Networks positions this release as bringing network and GPU infrastructure operations into one conversational interface. This Blog Summary is a fact-based summary of the vendor blog, and enterprise decision-makers evaluating AI-assisted operations may find the deployment, governance, and licensing details relevant.
