Telkomsel shifts gear to achieve highly autonomous networks by 2025

Among a group of telecom operators that also includes China Mobile, Orange and Zain, Indonesia’s Telkomsel is aiming to achieve TM Forum’s Level 4 in autonomous networks (AN) by 2025.

Telkomsel is confident that AN will play an important role in improving customer experience, increasing efficiency and enabling new services, but the path to network autonomy can be rocky. In this interview with Inform, Telkomsel CTO Indra Mardiatna explains how the company has had to adjust its course as it progresses.

From 2020 through to 2023, Telkomsel focused its AN effort principally on improving customer experience, as well as power saving measures. Successes to date, which according to Telkomsel, demonstrate the capabilities and advantages of its autonomous network since its initial launch in 2023, include:

• A twofold increase in mean time to repair (MTTR)
• A 15% reduction in customer complaints
• a net promoter score that is 27% above the industry average.

Other wins include the automation of network parameter checks. The company identified around 62% of total network parameters, , across its networks that a machine could daily (with average 70 thousands changes per day) resulting in a 3% to 5% increased in payload, as well as increased throughput.

Telkomsel’s goal is to increase its overall AN level as measured by TM Forum’s maturity model shown in the graphic below. Level 4 autonomous networks mark the transition between traditional automation of human-defined process behavior and autonomous behavior, where systems make decisions independent of humans.

“Telkomsel’s [AN] level was 1.8, which is behind where we want to be,” explains Mardiatna.

To meet its ambitious Level 4 objective by 2025, Telkomsel is using TM Forum’s AN Framework, and the company has brought in external experts to help. This enables Telkomsel to address what Mardiatna describes as its number one challenge, which is “talent – the number of people who understand how to implement automation.” Using the framework has also given the company greater access to relevant use cases.

As a result, Mardiatna expects “we will execute this year Level 4 for the main use cases, customer experience customer services and some of the qualities… and parameters.”

In addition, “in 2024 we are now trying to implement more network automation,” he adds. “We have also started with planning prediction /propagation of the number of subscribers and transit … and then we can make [traffic and performance] predictions [for individual] sites.

Focusing on use cases

Telkomsel has had to change its methodology since it began implementing its autonomous network strategy in 2020. Whereas the company initially took a tools-based approach to automation, it now works backwards from specific use cases.

“We changed our approach to based on use cases,” says Mardiatna, adding that “one of the most important advanced use cases the company is better data quality.”

This is because, like other network operators, Telkomsel, found that its initial forays into network automation highlighted the need for accurate, relevant data.

“In previous initial implementations we had a problem with data … which becomes a disaster for automation,” Mardiatna said.

By working backward from separate use cases Telkomsel will be able to focus its efforts on cleaning only the data it needs for a given deployment. This will save time, which will become increasingly essential as the company seeks to make greater use of AI.

“We want to develop and implement use cases for AI. And it can be done but it takes time. The process is important,” explains Mardiatna. “To have all the data to do things takes years. That is why our process based on use cases focuses on the specific data required [for a particular use case]. It must be continuous progress.”

Telkomsel is also working to tie together network performance and B2C, B2B and B2B2C customer requirements.

“We are looking at how we can exploit our infrastructures to work [for] the solution or vertical that we can offer to the customer,” explains Mardiatna. “If we can understand that behavior, then we can also open up some product solutions from the network that we can offer to them.”

In the longer term, Telkomsel has its eye on intent-based networks.

“Intent … is our dream, our aspiration, but it depends on the maturity of the infrastructure itself,” Mardiatna says. “Maybe some use cases we can get to that level by 2025, but in some very complex cases [we] require more power … and more data, and we need time.”

Indra Mardiatna will be speaking at DTW24-Ignite, which takes place 18-20 June in Copenhagen, as part of a panel discussion on ‘How To Progress Towards End-To-End Autonomous Networks.