No prizes for guessing that AI stole everyone’s attention at Digital Transformation World in Copenhagen this year. In every keynote and session, the two letter acronym left the lips of every single person that took a stage.
It is no wonder that AI dominated the show after so much fanfare already this year. The much hyped technology promises so particularly to birth new use cases that can save costs and generate fresh revenues.
Axiata Group CEO and Executive Director, Dr Hans Wijayasuriya said in his keynote that the telecoms industry can double its size to US$2 trillion dollars if it can “move from connectivity to a world where we sell end-to-end solutions”. He pointed to AI being a key technology to digitally transform operators into technology companies with this sort of scale and capability. He warned MNOs that do not jump on AI with haste, will be left behind to languish.
On the same panel, Indosat Ooredoo Hutchison President Director and CEO Vikram Sinha and Jio Platforms boss Kiran Thomas pointed to how AI can generate GDP growth for Indonesia and India respectively.
TM Forum CEO Nik Willetts pointed out in his speech that only a handful of MNOs are seeing bottom line increases from AI, but added many operators are still early in their digital transformation journeys.
Autonomous networks
In the telecoms space, automation of networks has been viewed as an eventuality to cut down costs but also create new services, and is a key use case for AI in telecoms.
Bradley Mead (pictured, right), Ericsson Head of Managed Network Services, said “autonomous networks is ultimately where we need to go” as an industry, but acknowledged it is a long journey but “starting it is vital”.
The ultimate goal for all sectors tapping into AI is to achieve ‘Level 5’ said Mead, which means full automation. This is considered the ‘holy grail’ for all automation projects, regardless of industry - Level 5 is akin to a fully driverless car with no need for a steering wheel, or a network that runs completely by itself with no engineer.
“Hopefully it [Level 5] will unlock new revenue streams for operators because ultimately, that's what we all need for the industry to be successful,” said Mead.
Sunil Gupta (pictured, left), Digital Nasional Berhad (DNB)’s Head of Digital Delivery, spoke about the challenges and opportunities of navigating autonomous networks in Malaysia’s singular 5G network.
Malaysia has a rather unique telecoms landscape, as the government decreed all six MNOs must equally share a single 5G network run by DNB, which was set up in 2021 to handle the deployment.
Gupta explained this is a Multi-Operator Core Network (MOCN) set-up, a key innovation in 5G that enables multiple operators to share common infrastructure while maintaining individual services and brands.
The idea was that this would cut down on expensive deployments, reduce rollout time and of course adapt to new technologies such as AI at speed. But in the time that Malaysia’s government has taken to strike deals with its operators, peers in Asia-Pacific such as Thailand and Indonesia have launched commercial 5G services.
The government had been steadfast in its strategy, betting there will be long-term returns from of its unique plan.
Gupta stressed that having operators share the same infrastructure and keeping their wealth of data private from each other was a key challenge.
To tackle this, DNB with its exclusive vendor Ericsson pushed through with the network being a “machine or data driven operation,” said Gupta. Through using AI technology, DNB was able to predict in real time where problems on the network would arise - one example was finding out why calls were being dropped in a particular location - and safeguard against it to maintain high levels of service.
Gupta revealed two main targets with its AI-boosted network: to provide enterprise network slicing and autonomous network energy management.
Levels of AI
At another stage, Telkomsel’s Director of Networks Indra Mardiatna (pictured, above) hailed the Indonesian operator’s successful implementation of autonomous network technology, highlighting benefits such as customer complaint handling, reduction of manual tasks, and improved NPS scores.
Mardiatna detailed how Telkomsel’s dive into automating its network has made processes smoother.
He pointed out the “legacy” way of handling a customer complaint had too many steps for subscribers, taking up valuable time. Now, Telkomsel customers only have to answer two questions, with analysis of their query coming “automatically from the network” including the root cause.
From this, Mardiatna claimed user complaint handling had improved by 88%, and that Telkomsel’s NPS score is 20% above the industry average.
Mardiatna noted that Christmas, Ramadan and New Year’s were three recent events when Telkomsel experienced heavy network traffic. These special days on the calendar used to require a lot of manual tasks to be completed to ensure the network ran smoothly.
“Previously when we experienced a push in threshold level, someone would have to execute manually an automation request, but now this can be done by machine,” said Mardiatna.
Telkomsel’s network was “90% automated” during the Muslim holy month of Ramadan for crowd management, as worshippers gathered daily at Mosques to pray. Mardiatna stated this worked six times better than having an engineer manually divert traffic to other parts of the network. This led to a 20% decrease in customer complaints, noted Mardiatna.
Looking ahead, Telkomsel is aiming to achieve level 4 AI automation in 2025. But Mardiatna acknowledged a few key challenges must be overcome to hit the target of having a “zero touch network” - a network that can heal and adjust itself based on the data it collects.
The first challenge that many operators have highlighted is quality of data and improving data validity. Mardiatna said due to poor data in the early days of using AI, the operator saw results that were a complete “disaster”.
Second, was cost for equipment such as GPUs to install into data centres to tap into AI software. Finally, third was reskilling staff, as the operator has “a lot of legacy telco people” who need convincing an autonomous network is the way forward.
“Despite all of these challenges, we believe autonomous networks will unlock all possibilities. We can achieve more and we believe also we move forward together,” concluded Mardiatna.
