Sri Lanka’s industry titans profess a pragmatic Artificial Intelligence path

Sri Lanka hosted its first-ever National Artificial Intelligence (AI) Expo and Conference on September 29–30. 

By Nishel Fernando
A high-powered panel at the National AI Expo and Conference 2025 revealed that for Sri Lanka’s leading conglomerates and tech firms, the era of theoretical AI is over. The focus is now squarely on practical implementation, measurable ROI, and overcoming foundational challenges in data, infrastructure, and skills to unlock the next wave of economic growth.
The air at the National AI Expo and Conference 2025 in Colombo was thick with discussions of large language models, neural networks, and sentient machines. Yet, in a packed auditorium, a panel of the nation’s most influential technology leaders brought the conversation firmly back to earth.
Moderated by Arun Pai, Senior Principal at Analysys Mason, the session on “Sector-Specific AI Applications” moved beyond the hype to dissect the real-world adoption of Artificial Intelligence across Sri Lanka’s core economic sectors: telecommunications, finance, retail, and manufacturing.
The panel featured a formidable lineup: Dr. Ramesh Shanmuganathan, Group CIO of John Keells Holdings; Asanga Ranasinghe, CTIO of Hutchison Telecommunications Lanka; Dr. Harsha Subasinghe, the founder and CEO of CodeGen; and Dr. Raja Mansukhani, Chief Strategy, Technology and Transformation Officer at Kambiva Technology. Their message was that AI is not a magic bullet, but a powerful tool that requires strategy, robust data foundations, and a fundamental shift in corporate mindset to deliver tangible value.
The Strategy: ROI before AI
For a conglomerate as vast and varied as John Keells Holdings, with interests spanning from supermarkets to banking, the temptation to sprinkle AI across every business unit could be overwhelming. However, Dr. Shanmuganathan outlined a far more disciplined approach, rooted in years of experimentation dating back to 2014.
“We look at two things when we look at AI: strategy and automation,” Dr. Shanmuganathan explained. The crucial question, he noted, is not about chasing the most advanced AI models, but about identifying business problems where existing data sets can be used to generate a clear Return on Investment (ROI). “It’s about ROI. It’s not about cost. Even if we invest more in AI, if we know the ROI, then obviously we can make that call.”
This pragmatic lens has focused John Keells’ efforts on high-impact areas. Dr. Shanmuganathan cited several use cases currently in motion: optimizing supply chain efficiency, enhancing marketing effectiveness, and even predicting the precise cash requirements for the group’s ATM network to avoid tying up capital unnecessarily. “Tell me what the optimal discount to show me,” he said, illustrating how AI can drive data-led decisions at the consumer level. This focus on tangible outcomes over technological novelty represents a maturing of the corporate approach to AI in Sri Lanka.
The Foundation: Data as the Bedrock and the Bottleneck
Every panelist agreed that a successful AI strategy is built on a single, non-negotiable foundation, which is data. Without clean, accessible, and well-governed data, the most sophisticated algorithms are useless.
Asanga Ranasinghe detailed the critical first steps his organization took. “Being a Telco, we have so much of data,” he stated. “As a strategy, what we had to do was to make sure that we have a data lake where all this data is fed, so we can use this to train the AI models.”
However, amassing data is only half the battle. The panelists highlighted the immense challenges of governance, security, and privacy.
 «We also need to manage data privacy,» asserted Dr. Shanmuganathan, emphasizing that John Keells does not share personally identifiable information between its different industry verticals. This commitment led the group to establish a robust “zero trust” security architecture, classifying data and tightening identity management before scaling its AI initiatives. “It’s important for all of us to look at data more closely,” he urged.
Ranasinghe echoed these concerns, noting the complexities of data sovereignty. “We have certain restrictions; we cannot put your PII (Personally Identifiable Information) data over there [on public clouds]. So, what we can do on-premise versus what we can do on the cloud, this decision has to be made.” This delicate balance between leveraging cloud-based AI platforms and protecting sensitive customer data remains a central challenge for Sri Lankan enterprises.
The Application: AI in Action Across Industries
With strategic frameworks and data foundations in place, Sri Lankan companies are deploying AI in increasingly sophisticated ways.
For telcos such as Hutch, AI is revolutionizing both customer experience and core network operations. Ranasinghe categorized their approach into “low-hanging fruits” and more complex, long-term projects.
Immediate wins include AI-powered chatbots for customer service and internal support, alongside analytics dashboards that allow management to ask natural language questions like, “What is today’s revenue and how is the forecast going to be for the next seven days?”
The real game-changer, however, lies in network operations.
“For me as a CTIO, the most important thing is predictive analytics and network optimization,” Ranasinghe said. 
Hutch is working on AI models that can predict network congestion or site failures before they impact customers. Furthermore, AI is set to automate the painstaking process of network optimization. Instead of engineers performing manual “drive tests” to adjust cell tower parameters, AI can do it on the fly. “In the coming future, network optimization will be very efficient, fast, and less costly,” he predicted.
Dr. Harsha Subasinghe of CodeGen provided a glimpse into the future, where AI is not just a tool to be applied but an integral part of product creation. His company is a leader in building complex AI systems from scratch.
 «We use vision-based navigation as well as LiDAR,” he said, describing their work in occupational robotics and autonomous systems, including for the acclaimed Vega electric supercar project.
Dr. Subasinghe also demystified the technology, emphasizing that AI will primarily replace monotonous tasks, freeing up human potential. 
“AI is going to replace all the boring things that we do,” he declared. He envisions a world where AI democratizes expertise, allowing a large language model to provide instant access to the world’s knowledge, whether for a student, a lawyer, or even a doctor diagnosing symptoms based on biomarkers.
For a technology company such as Kambiva, the challenge is twofold: implementing AI internally and enabling clients to do the same. Dr. Raja Mansukhani shared the structured four-block framework his company developed to navigate this.
Becoming an AI-First Company: Embedding AI in product functionality, engineering, and operations.
Engineering Excellence: Using AI to augment developers and focus human talent on strategic architecture and design.
AI-Powered Operations: Proactively monitoring production services to enhance customer experience.
Cybersecurity & Governance: Using AI to bolster security while ensuring the ethical and secure governance of AI systems themselves.
This framework, he explained, helps cut through the noise and focus on value creation. “It was very important for us to structure a framework in terms of how can we focus on value creation and capture for our clients,” Dr. Mansukhani noted.
The Enablers: Infrastructure, Connectivity, and the Role of Telcos
A recurring theme was that the nationwide adoption of AI is critically dependent on the underlying digital infrastructure. Here, the role of telecommunications providers extends far beyond being mere users of AI.
“Our role will be to provide connectivity,” stated Asanga Ranasinghe, envisioning telcos as the primary enablers of an AI-powered economy. He stressed the need for a multi-technology strategy—combining fiber, wireless, and emerging low-earth-orbit (LEO) satellite services—to ensure ubiquitous coverage across Sri Lanka.
The impending arrival of 5G is set to be a major catalyst. “When 5G comes... and when the AI machines are connecting to the network, there will be a lot of upload of data. The 4G network cannot handle it,” he explained. The high bandwidth and low latency of 5G are essential for the massive machine-to-machine communication that industrial AI applications will require. Beyond connectivity, Ranasinghe also pointed to discussions around investing in crucial data center and GPU infrastructure locally, a vital component for training and running powerful AI models within the nation’s borders.
The Human Element: Upskilling a Nation for the AI Era
Perhaps the most critical question was what skills does the Sri Lankan workforce need to stay relevant in an AI-powered future? The panel’s answer was not just about coding, but about a broader, more holistic understanding of the AI ecosystem.
Dr. Raja Mansukhani outlined three urgent priorities:
AI Awareness: Moving beyond casual use of tools like ChatGPT to formally understand the AI ecosystem—how it works, its potential, and its limitations.
Continuous Self-Learning: Recognizing that the pace of technological change is accelerating, making the ability to constantly learn and adapt a paramount skill.
Understanding AI Governance: Appreciating the critical importance of data privacy, security, and the ethical implications of AI, as these will be central to building trust and ensuring responsible deployment.
Dr. Harsha Subasinghe offered a reassuring, albeit challenging, vision. He argued that current AI is not sentient—”it doesn’t have emotions, it can’t think”—and that humans need not fear being replaced in an existential sense. He took the idea a step further by looking at the frontiers of research into Synthetic Intelligence, a future concept distinct from today’s silicon-based AI. He described emerging work on creating intelligence using biological components. “You can have 85 billion neurons running on 20 watts of power,” he noted, contrasting the human brain’s incredible efficiency with silicon-based models that can require megawatts of power.
This path, he mused, raises profound questions. “When you go through Synthetic Intelligence, whether it can... become sentient or not is a question, because it’s human brain cells.” However, after touching upon this fascinating frontier, he brought the discussion back to the present reality. For now, he concluded, the focus should not be on such futuristic possibilities. “The jobs that we are doing right now have to be changed,” he warned. “That is one thing that we need to look at, not to worry about the AI.”
As the panel concluded, the path forward for Sri Lanka became clearer. The nation’s corporate leaders are not waiting for a far-off technological singularity. They are on the ground today, building the data pipelines, investing in infrastructure, and strategically deploying AI to solve real-world business problems, ensuring that Sri Lanka is not just a consumer, but an active participant in the Artificial Intelligence revolution.