Page 359 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

AI Copilot to simplify network operations:

The integration of an AI Copilot powered by Natural Language to SQL (NL2SQL) technology
enables users to perform tasks such as fault diagnoses, configuration changes, and policy
recommendations using natural language dialogs. This significantly reduces operational 4.3 - 5G
complexity and facilitates the transition from expert-driven to AI-powered network management.

The telecom foundation model is trained on the cloud, and then model-based applications
are deployed in MTN. With the AI-based system, the packet loss rate and network delay
are significantly reduced, and network utilization is greatly improved. This effectively solves
network congestion problems caused by power rationing and line interruption, improves user
experience, and further eliminates the digital divide.

Partners
• Huawei Technologies [6]

2�2 Benefits of the use case

The use case deploys AI to enhance quality, reliable and resilient internet infrastructure which
can withstand challenges, notably power outage.

Next-generation transport network technologies and agents play a crucial role in boosting
network resource efficiency and cutting down device power consumption. AI agents empower
intelligent management and optimization of network devices, minimizing the environmental
footprint of network operations. This not only advances eco-friendly development in the
telecom industry but also strengthens network infrastructure resilience, ensuring uninterrupted
and reliable network communication services for all, increase the accessibility to internet in the
region and effectively bridging the digital divide.

2�3 Future Work

Continuously innovate network optimization oriented to the AI WAN. This includes achieving
automatic network analysis and optimization based on KPI targets, as well as delivering regular
network insight reports that utilize analysis capabilities such as traffic prediction and suppressed
traffic visualization. These efforts aim to enhance service experience assurance capabilities and
meet the network access requirements of a broader public.

3 Use Case Requirements

• REQ-01: It is critical to enable high-precision and high-accuracy SLA measurement for
network-wide E2E traffic suppression visualization.
• REQ-02: It is critical to support millions of tunnel management and minute-level network-
wide path computation capabilities for large-scale network management.
• REQ-03: It is critical to enable performance instances for network objects to obtain the
required performance indicators for network analysis.
• REQ-04: It is critical to provide the appropriate deployment environment for the system,
such as servers and computing cards.

323