Telkomsel to implement AI for 4G network

26 November 2024

Telkomsel has teamed with ZTE to implement an AI-powered network solution in its 4G network to optimise base station performance after completing tests in Makassar and Kendari.

The ‘self-adaptive feedback solution’ promises to optimize network performance without the need for additional hardware. The solution leverages AI and machine learning to automatically adjust network parameters such as speed and power control, for data-intensive apps such as video streaming and gaming.

Telkomsel has been testing the solution in more than 90 sites on its live network in Makassar and Kendari, covering 300,000 users. The results saw video buffering reduced by 15%, download speeds increased by 11%, web page loading times improved by almost 30%, and latency for gaming was reduced by 47%.

“By integrating artificial intelligence into the network, we are not only increasing efficiency, but also answering the increasing need to access high-quality digital content,” said Richard Liang, president director of ZTE Indonesia.

Telkomsel’s planning and transformation director Wong Soon Nam said that the self-adaptive feedback test is part of the telco’s ‘Hyper AI’ strategy to utilise end-to-end AI and ML to improve the customer experience.

The self-adaptive feedback solution also improves energy efficiency, which translates into lower opex costs. By automatically switching the base station to low-power mode when traffic is low, the solution increased power efficiency by about 15% while lowering energy consumption by 8%.

With the tests in Makassar and Kendari complete, Telkomsel will deploy the solution across the rest of its network in Indonesia.