TECH NEWS – Artificial intelligence is no longer merely an experimental technology in telecommunications. It is increasingly being used for network planning, fault detection, operations, and energy management, and CETIN Hungary says several of these applications are already part of its domestic infrastructure while additional use cases remain under evaluation.

CETIN Hungary recently completed a modernization of its network infrastructure covering more than 4,080 base stations, an investment exceeding HUF 120 billion, and is now adding more automated and AI-assisted capabilities to that modernized network. The company is examining or already using these tools in areas ranging from network planning and operations to power consumption. As part of a recent professional program focused on AI in telecommunications, CETIN also worked with AI expert and digital transformation executive Balázs Koltai to assess where the technology can have a practical role in telecom infrastructure.

Koltai argues that the industry’s focus is increasingly moving away from traditional capacity expansion alone and toward network intelligence. “AI-supported planning and operations make the organization more efficient and free specialists from routine tasks,” he said. Longer term, he expects some base stations to take on Edge AI functions, allowing certain computing workloads to be processed closer to users instead of being sent to distant data centers. That approach could be particularly relevant for applications requiring very low latency.

AI Is Moving From Network Planning to Automated Recovery

One of the most straightforward applications is planning where network capacity should be expanded. AI can analyze aggregated telemetry data alongside user feedback to estimate more accurately whether a particular area requires a new base station, additional capacity at an existing site, or another type of network adjustment. CETIN is currently evaluating how much this cell-level approach can accelerate and deepen the planning process, and whether the same investment can produce a measurably better subscriber experience.

AI also plays a role after the network is already operating. CETIN continuously assesses it through thousands of key performance indicators, which can be analyzed together with forecasts for future traffic. The system can, for example, identify the 100 worst-performing base stations and assist engineers in designing a more appropriate radio configuration. Predictive analysis may flag certain anomalies before they develop into larger service problems, while some issues can be corrected automatically. CETIN describes this as a form of network “self-healing,” although more substantial interventions still depend on engineering decisions.

Generative AI is also being used in less visible operational workflows. Engineers create detailed documentation and work tickets for detected network problems before deciding what intervention should take place. CETIN uses generative AI to review those tickets, identify missing information, and provide technical feedback. In this case, the technology is not independently deciding how the network should be modified, but supporting the preparation and quality-control process surrounding the work.

Energy consumption is another major area of interest. By continuously examining consumption patterns, AI can identify behavior that differs from expectations and may point to hidden faults or other anomalies. Network elements can also be controlled dynamically according to predicted traffic: during low-demand periods, unnecessary capacity may be temporarily switched off or placed into a power-saving “smart sleep” state before being restored when demand increases. CETIN says intelligent network control and existing power-saving functions currently save roughly 3 million kWh of electricity per year, which the company compares with the annual electricity demand of 1,295 average Hungarian households. According to CETIN’s estimate, targeted AI assistance could increase the amount of electricity saved by a further roughly 20 percent.

Source: CETIN Hungary / Noguchi