(Source: Christoph Theisinger on Unsplash)

Sri Lanka Telecom (SLT) wants to become an AI-native telco, a goal one of its senior executives acknowledges is far from easy. "It has become a strategic priority for us to build an AI-native telco, although it is a very ambitious target," Ishari Siriwardane, deputy chief transformation officer at SLT, told Light Reading during an interview on the sidelines of FutureNet 2026.

Few service providers carry as much history as SLT. The company began as a government department in 1858, went through privatization with Japan's NTT as a strategic investor, and is now a blue-chip name on the Colombo Stock Exchange, with the government still holding 52%. It competes with Dialog Axiata and Hutch Sri Lanka.

SLT started about four years ago not with AI but with what it calls a digital maturity assessment. "There were about 20 plus initiatives that we started off on our digital transformation program," she says. That led to a dedicated digital transformation division, now housed within the IT group. The company has also formed a Center of Excellence team for AI and has appointed a head of data and AI, revealed Siriwardane.

Where AI is already working

In customer service, chatbots and assistants now face customers, and there are also AI agents at the back end to prevent staff from having to log into multiple systems when handling a complaint. "Now we have AI agents being built at the back end so that they can fetch this information quite quickly and then support the customer," Siriwardane said, adding "so the cycle time that we take to resolve an issue for the customer has improved."

Credit control is next, with personalized collection actions made possible with AI. It can assess factors like customer behavior and payment history, as well as what triggers them to make a payment, she stated.

On the network side, AI is helping with fault identification and resolution. "We have a very fragmented and heterogeneous network, which is a challenge," she says. AI is helping with fault identification and resolution, particularly in wireless network, but it is "not gone up to the autonomous level." The task now is scale: "So that we can make it the default setting in the organization so that everybody uses it."

The hardest problems aren't technical

The company's history is both an asset and a drag. The first issue is that the average employee at SLT is more than 45 years old. "Changing the mindset of the people, getting them to adapt to AI and also sustain that change has become a challenge." The task is "moving away from this fear factor that AI is going to take my job kind of a scenario we need to deal with. So we need to have that cultural shift within the organization."

Another issue is legacy infrastructure, which makes it difficult to adopt and implement new solutions. "Running AI on top of the legacy infrastructure is quite challenging," she says, and modernization needs capital when margins are thin. "We don't want to burn our fingers as well by spending a lot of capex," she says. There is a balance to strike between investing in modernization while maintaining profitability.

The third challenge is state ownership. "We are also bound by government processes and procedures," said Siriwardane, since the company is 52% government owned. She added this means dealing with problems like red tape.

5G, but only where it pays

SLT launched 5G last year and like other service providers it is struggling with monetization. The new commercial angle is AI tokens. "We are looking at new products and solutions where we can bundle AI tokens with 5G connectivity," said Siriwardane.

Since 5G use cases are limited, SLT has adopted a somewhat different approach with deployed 5G. It launched 5G last year and had decided against a pan-country rollout in Sri Lanka, an island country with a population of 22 million in 2025.

"It [5G monetization] is a challenge for us as well," she said. "Since we are a small country, we don't want to deploy 5G all across the country." Therefore 4G remains the mainstream network, while 5G would be used in strategic areas, such as places with high business and economic activity like airports or stadiums.

Training 3,500 people

"Right now we have about 3,500 plus staff which we plan to train," Siriwardane says. "So 50% of this is already trained and we are targeting to complete the training […] by the end of the year.”

Commenting on AI tools, Siriwardane noted that they "have provided Microsoft Copilot licenses for everybody." They have made available different versions of MS Copilot depending on the profile of the employee.

At the same time, SLT is also evaluating cheaper open source models for internal use to bring down cost. "We are cost conscious," she says, with token costs rising, and "we want to make sure that it doesn't hit our balance sheet."