Digital network infrastructure provider Zayo is extending its network as a service (NaaS) model so approved artificial intelligence (AI) agents can interact directly with the network.
Zayo said its Agentic Networking system advances NaaS from self-service network control to agent-enabled networking, where teams define objectives and controls while AI helps coordinate and execute network tasks.
Outlining the importance and context of the launch, Zayo noted that the first generation of NaaS gave enterprises software-based visibility and control over network infrastructure. Now, with the new system, Zayo said Agentic Networking changes how that control is consumed by making network capabilities accessible in the AI systems enterprises are already using to coordinate work across applications, clouds and other systems.
The end result is that agents can investigate performance issues and take authorised actions within enterprise-set boundaries, evolving NaaS from human self-service towards governed, agent-driven network operations.
“Enterprise AI will increasingly be measured by what it can responsibly accomplish across the systems that run the business,” said Max Clauson, senior vice-president of network connectivity at Zayo. “The future of networking is not another portal. It’s making network capabilities available within the AI systems enterprises use to run their business. Agentic Networking gives organisations a path toward that future, where AI can take on more operational work while people maintain authority over the infrastructure underneath it.”
Zayo claims to have introduced a new approach to software-defined networking with DynamicLink in September 2025. Now, with Agentic Network, as DynamicLink reaches its first year in production, Zayo says it’s introducing the industry’s first production Model Context Protocol (MCP) server built for networking, progressing DynamicLink from self-service network control to Agentic Networking. This is said to give approved AI agents a governed way to understand network conditions, access network tools and take authorised action, creating a capability that the company says has not previously existed.
“This combination creates a way for approved AI agents to understand what is happening across the network, access network tools and execute authorised actions,” said Zayo chief revenue officer for enterprise, Chris Ranalli. “Instead of simply asking AI questions about the network, enterprises can put AI to work on it.”
Agentic Networking brings network context and authorised network actions into the AI workflows enterprises are already building. Agents can use product and network knowledge resources to understand network conditions, while retrieval-augmented generation (RAG) helps ground interactions in the DynamicLink knowledge base. Organisations define the information, tools and actions available to agents, while preserving automation, governance and human oversight.
Potential user workflows indicated by Zayo include establishing cloud-to-cloud connectivity; investigating network performance; coordinating multi-cloud connectivity; responding to unusual network activity; executing approved network changes; and extending networking into enterprise AI workflows.
That is, firms can coordinate the steps required to connect approved cloud environments without engineers manually moving between multiple systems. They can also gather and correlate context across network, cloud and application environments to help identify the source of an issue and recommend next steps.
In a multi-cloud environment, the system is built to help teams understand and manage connectivity requirements as workloads move across cloud environments. In addition, it can surface relevant network context, identify potential issues and recommend actions for an engineer to review. This also has the possibility to invoke authorised tools to carry out network or security actions within enterprise-defined controls.
Zayo stressed that the overall result in deploying the system will be less time in moving between systems and coordinating routine tasks and more time focused on higher-value decisions. Zayo believes this efficiency is especially valuable for industries such as retail, financial services, healthcare, and media and entertainment organisations managing increasingly data-intensive environments.
