Google Cloud and Nokia are adding agentic AI to telecom networks.The system exposes network APIs so AI agents can manage connectivity.
Telecom networks have long been managed through scripts, manual rules, and layers of software tools. A new effort between Google Cloud and Nokia points to a different model: networks that can respond to goals set by software agents rather than detailed human commands.
The two companies are working together to bring agent-based artificial intelligence into Nokia’s Network as Code platform. The idea is to allow software agents to interact with telecom network functions through APIs, rather than relying on telecom engineers to configure each step manually. The approach could shift how telecom infrastructure is used by both operators and enterprises.
Turning telecom networks into programmable systems
Nokia introduced its Network as Code platform to expose telecom capabilities through APIs. These APIs allow developers to build applications that interact directly with network features such as connectivity quality, device location checks, or network slicing. According to Nokia, the platform is designed to simplify access to functions that were traditionally locked inside telecom infrastructure.
The collaboration with Google Cloud adds an AI layer to this system. Instead of developers calling APIs step by step, an AI agent can interpret an instruction and determine which network functions to use. Google Cloud describes agentic AI as software that can observe a system, make decisions, and carry out actions toward a goal.
In practice, that means an AI agent could request certain network conditions for an application, such as stable connectivity for a delivery drone or priority bandwidth for a remote industrial machine. The system would then adjust network resources to meet that goal.
This kind of automation is being explored across the telecom sector as networks become more complex with the spread of 5G, edge computing, and connected devices.
Why networks are becoming harder to manage
Telecom operators already rely on automation tools to keep networks running. Yet the scale of modern networks has grown far beyond earlier systems.
5G networks introduce new features such as network slicing, where operators can allocate parts of the network for specific uses like industrial control or private enterprise connectivity. These capabilities add flexibility, but they also increase operational complexity.
Industry analysts say this complexity is pushing operators toward new forms of automation. Industry forecasts suggest that mobile networks will need to support billions of connected devices by the end of the decade, making manual network management increasingly difficult.
Agent-based AI is one approach to handling this scale. Instead of engineers setting every rule, software agents could manage network behaviour within defined limits.
How agentic AI fits into telecom operations
The collaboration between Google Cloud and Nokia centres on three layers.
The first layer is the network API system. Nokia’s platform exposes network features through standard APIs. Developers can use these APIs to request network capabilities without needing deep telecom knowledge.
The second layer is the AI system. Google Cloud’s AI models interpret instructions and decide which network functions should be used to achieve a goal.
The third layer involves software agents interacting with each other. These agents can monitor conditions, adjust network settings, and coordinate actions automatically.
Google Cloud says this type of setup may allow applications to interact with telecom networks through natural-language instructions rather than low-level telecom commands.
The approach reflects a broader shift in how AI is used in infrastructure. Instead of analysing data after the fact, AI systems are beginning to act as operators that can take actions inside complex systems.
Potential uses beyond telecom operators
Network APIs have attracted interest from industries that rely on reliable connectivity.
For example, a logistics system could request stable network coverage for autonomous vehicles moving through a city. A manufacturing system could request low-latency connections for factory robots. A streaming service might request extra bandwidth during large live events.
These kinds of use cases depend on the ability to control network behaviour in real time.
Nokia says its Network as Code ecosystem already includes more than 75 partners, including telecom operators such as Deutsche Telekom, Telefónica, and Vodafone. The ecosystem is meant to give developers access to network capabilities across different operators.
Challenges that remain
While the concept of AI-managed networks is gaining attention, practical deployment still faces several hurdles.
Telecom infrastructure is heavily regulated and must meet strict reliability standards. Allowing software agents to modify network behaviour introduces questions about oversight and safety.
Operators must also ensure that AI systems do not create unpredictable network behaviour or security risks.
Another issue is interoperability. Telecom networks are built from equipment supplied by many vendors. For AI agents to manage these networks effectively, they must work across different technologies and standards.
Industry groups such as the GSMA and the 3rd Generation Partnership Project are already working on frameworks for network APIs and automation to help address these challenges.
How agentic AI could change telecom networks
The effort between Google Cloud and Nokia reflects a larger change in how telecom networks may operate in the future.
For decades, telecom infrastructure functioned mainly as a transport system for data. Network features were controlled by operators and rarely exposed to outside developers.
Network APIs and agent-based AI suggest a different model. In this model, networks act more like programmable platforms where applications can request specific network behaviour.
Whether that shift becomes widespread will depend on how telecom operators balance automation, reliability, and security.
What is clear is that as networks continue to grow in scale and complexity, the industry is searching for new ways to manage them. Agent-based AI may become one of the tools used to handle that challenge.
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.
AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.