Newly launched MCP server provides structured, high-level access and orchestration capabilities, while Alkira’s NIA copilot guides network administrators through configuration and troubleshooting.
Network-as-a-service (NaaS) startup Alkira has spent the past year building momentum in the enterprise networking market, growing revenue approximately 60% and landing its largest customer to date, a Fortune 10 company. That growth has coincided with a wave of network transformation projects driven by shifts in traffic patterns and infrastructure architecture. A year ago, the company released zero-trust network access (ZTNA) functionality, which helped to improve security. Now Alkira has turned its focus to the same topic that all IT vendors are talking about in 2025: AI.
Alkira’s AI focus involves three core areas: networking for AI workloads, AI for networking operations, and internal AI use across development and support functions. Today, it’s launching a Model Context Protocol (MCP) Server and Alkira NIA (Network Infrastructure Assistant). MCP Server enables agentic AI systems to query Alkira’s control plane alongside other backend systems like ticketing platforms and monitoring tools. NIA functions as a conversational co-pilot embedded directly in the Alkira platform for troubleshooting, configuration and monitoring tasks.
“The whole goal here is to automate more and more and make our operators’ lives simple and easy,” Manan Shah, senior vice president of products at Alkira, told Network World.
MCP Server: Unifying interfaces and automation
Over the course of the last year, MCP server technology has increasingly become a defacto standard across all types of applications. Functionally, MCP is a JSON API layer that enables AI agents to access and in some cases control a service.
The Alkira MCP Server sits as a layer above the company’s existing API infrastructure. While that positioning might suggest a simple API wrapper, Shah emphasized the system provides substantially different output.
“It’s not just an API wrapper, right?” Shah said. “It’s not just about accessing the same APIs in a different way, but it’s also about getting more information, more digestible information, more humanly readable information in the network, from the network.”
The practical difference spans a number of areas including the format of responses. Rather than returning raw JSON that requires parsing and interpretation, the MCP Server can deliver tabular summaries. An operator or AI agent can request a deployment overview and receive structured data showing region counts, segment configurations, connector types, operational status and active alerts.
This preprocessing serves the primary use case: integration with agentic AI systems that need to query multiple backend systems simultaneously.
Shah noted that an AI agent could be talking to multiple MCP servers in support of an overall service. For example, the AI agent could be talking to Jira, ServiceNow and a monitoring service to get information and take action.
The MCP Server inherits Alkira’s role-based access controls and security frameworks. Users can only access data they’re authorized to see, maintaining the existing governance model.
NIA copilot: Bringing conversational troubleshooting to networking
The MCP server is an integration and automation platform intended primarily for agents, APIs, scripts, or other systems, providing structured, high-level access and orchestration capabilities.
For end-users, Alkira’s NIA copilot extends the utility of AI into everyday network operations. With a chat-style interface, NIA guides users through configuration, troubleshooting and gathering documentation.
NIA not only analyzes and identifies network issues but walks users step-by-step through diagnosis and remediation. For example, if connectivity fails between two IP addresses, NIA traces the path, checks the connector’s health, gathers logs and recommends specific troubleshooting action.
Recognizing the importance of determinism and reliability in networking, Alkira places a premium on explainability. The system provides transparency by detailing its diagnostic process, allowing users to trust, but also verify, AI-generated recommendations.
“We not only provide the answers, but we also provide the steps of how we’ve gone through that,” Shah said.
NIA will also enable Alkira users to more easily handle complex workflows with fewer steps. Shah noted that while existing users have generally liked Alkira’s user interface, it’s somewhat geared for certain types of workflows.
“NIA will essentially be your copilot to do any dynamic workflow that you want,” he said. “Let’s say you want to do five things across five different parts of the application. Now it can essentially guide you through that.”
