In Espoo, Finland, Nokia announced a step toward network operations that trusted fewer humans and more algorithms. The company introduced an agentic AI framework within its Network Services Platform (NSP), a comprehensive management system for multi-vendor IP networks. The framework lets network operators deploy AI agents that can reason over real network data and take guided actions—but only within policy and security boundaries they define themselves.
The timing matters. As AI traffic grows and networks expand in complexity, operators face mounting pressure to improve efficiency and reliability without losing control. Many have hesitated to embrace AI-driven automation, wary of unpredictable outcomes in production environments where mistakes can cascade across entire systems. Nokia's solution directly addresses that caution by embedding agentic AI capabilities into NSP, the platform already serving as the authoritative controller for IP networks.
The framework grounds AI agents in an accurate, continuously updated view of network reality: topology, protocol behavior, configuration state, service relationships, and recent changes. Rather than reasoning from fragmented or inferred data, agents work from network truth. They operate within operator-defined intent, policies, and access controls. The framework also enables communication with external agents via AI-based protocols such as Model-context protocol (MCP) across multi-vendor, multi-domain networks—a capability that empowers operators on their journey toward fully autonomous networks.
Grant Lenahan, Partner and Principal Analyst at Appledore Research, frames the approach: "Quality data and ontological relationships are proving far more important than specific AI models for efficient and accurate AI reasoning. Nokia's NSP embraces this approach with extensive AI-native infrastructure built on trusted data and operating norms, providing a solid and secure foundation for a myriad of AI use cases."
The first practical application is the Nokia AI-driven Troubleshooting Agent, built on this new framework. It accelerates root-cause analysis by helping operators identify what went wrong faster, reduces operational noise—the flood of alerts that obscure real problems—and transforms complex IP issues into guided, explainable workflows. For network teams, this means faster fault resolution, improved service reliability, and reduced risk of prolonged or cascading outages. End-users benefit through better experiences without increased operational risk.
Sasa Nijemcevic, Vice President and General Manager of Nokia's IP Network Automation software unit, emphasizes the philosophy: "Trust remains the deciding factor. We are enhancing NSP with AI agents built on an agent framework in a way that respects how networks are actually operated. This is an incremental, pragmatic step toward AI-native networks."
For operators, the framework provides a flexible foundation to introduce multiple AI use cases over time without creating isolated solutions. They can start with focused, high-confidence scenarios—like troubleshooting—and gradually expand AI's role as trust builds. A shared framework enforces consistent governance and operational controls across all use cases, preventing the messy proliferation of point solutions that often plague enterprise technology adoption.
The enhancement will be commercially available by the end of 2026. It represents Nokia's commitment to enabling trusted, AI-native network operations at scale, translating AI innovation into measurable operational outcomes rather than theoretical possibilities.