The Internet of Agents (IoA) reflects one of the most significant market shifts in the telecom industry. As physical network assets from cell towers to fibre become more programmable, the industry is embracing an agentic AI revolution. This is underscored by network operators deploying billions of autonomous agents within the network and operations by 2030.
Each agent acts independently to perceive, plan, act, and collaborate on behalf of network provider to personalise customer experience, improve network performance and accelerate autonomous and context-aware infrastructure. There are three major domains where agentic systems are helping the carriers to change their business.
Driving monetisation and NPS via personalisation
For the service layer, AI agents will allow customers to benefit from a strong user experience through the possibility to simplify app usage, customise settings, and enable voice command-driven services such as online shopping, booking travel, or even supporting daily tasks.
Given the importance of customer privacy, personalisation is also balanced with carrier-grade security, data protection, and higher levels of identity assurance. These agents can extend from mobile devices to home networking and across to multiple B2B cases (e.g., CCTV, video analytics, smart POS). Operators can also unlock new types of services, such as AI calling, improving voice clarity and MOS scores, and creating new revenue streams from a wide range of multimedia applications. Carriers that invest in intuitive apps, seamless service journeys, and tailored experiences tend to see immediate improvements in NPS scores, ARPU and monetisation through service creation.
From lag to brag: accelerating network performance
Agentic AI transforms network performance by shifting from the industry’s reactive, human-driven approaches to proactive, autonomous, goal-oriented systems. AI helps operators in areas such as fault management (e.g., anomaly detection, root cause analysis, and predictive maintenance) significantly faster than humans. Agents are equally able to perform autonomous diagnosis, isolation and resolution, enabling carriers to achieve near-zero touch incident resolution.
This reduces network degradation and the risk of outages. Huawei estimates that the use of these technologies can reduce the mean time to repair (MTTR) by at least 20, thereby improving service assurance, network resiliency and uptime. The vendor has reported having more than 20 use cases in operations across 120 CSP networks.
Embedding agentic AI into network elements and hardware
From codes to core networks are starting to think for themselves. Like other environments, agentic AI allows network elements (optical, WAN, or fixed access networks) to allocate and adjust resources dynamically based on real-world conditions. This can be RAN adjusting to increase data throughput from limited spectrum during a peak period, or directing base station resources to turn on, direct or amplify resources to the users and devices rather than broadcast in multiple directions. Other network elements include better ability to self-provision, self-heal, increase capacity and dynamic re-routing.
Agentic AI is making carrier networks adapt to changing conditions like traffic patterns, user demands, service priorities, and traffic flows. It is also allowing the industry to find solutions for physical challenges in delivering better coverage, higher capacity, lower latency, higher levels of reliability and throughput. This is also extending to intent-based routing by setting the network parameters to meet the end-user demands or application requirements. These are only possible through AI and automation due to the scale and complexity of the networks.
At MWC 2026, Huawei has provided a roadmap of how it plans to leverage AI/ML algorithms to improve network control policies to improve performance, energy efficiency, scale and user experiences. This moves the industry from one-size-fits-all approaches to more dynamic environments.
