Software will eat the world,” declared Marc Andreessen in 2011, as the VC firm he co-founded invested in a slew of software startups that would change the world. Six years later, Jensen Huang, CEO of Nvidia, offered an update: “Software is eating the world, but AI is eating software.” These words mark the shift that’s reshaping the global software industry. Today, Generative AI (GenAI) tools like ChatGPT, Copilot, Gemini, and CodeWhisperer have automated much of the raw coding that once occupied armies of junior developers.
That disruption is already visible in Sri Lanka. “Jobs are going to go. Let’s not shy away from that,” says Hariharan Padmanaban, Country Head – Sri Lanka of Hexaware Technologies, a global technology and business process services company. Shanil Fernando, Chief Technology and AI Officer of Cut+Dry, a software product company of 120 people, agrees: “We’ve stopped hiring. The main reason is significant productivity gains from AI-enabled tools. If there are strong AI-enabled engineers, we would hire them. But they’re quite rare.” Tools such as Cursor AI, he adds, already handle much of the work once done by junior developers.
The premium skill is no longer in programming syntax but in the ability to clearly define a problem, guide the tool, and validate its output in solving it. The effect on the IT job market is immediate. Internships are scarce, entry-level positions are disappearing, and clients are pressing IT firms to deliver with leaner, AI-augmented teams. Thus, eliminating once established pathways for fresh graduates to enter the industry. For Sri Lanka, the shock is sharper still. This disruption comes as the industry is recovering from the migration shock during the economic crisis that drained its middle layer of experienced technology talent.
How Is Generative AI Changing Software Development?
With GenAI automating large parts of raw coding, it’s reshaping the way software firms work. “A domain-savvy person using GenAI can do the work of four or five coders,” says Hariharan, who leads a team of 30 professionals in Sri Lanka as part of Hexaware’s global operations. He continues, “If a good programmer writes 1,000 lines a day, an AI tool can write 5,000. But only if the person prompting it knows exactly what they’re trying to solve.” Meanwhile, at Calcey, a boutique technology consulting and software product engineering services provider with 200+ employees, AI use is now mandatory, which Mangala Karunaratne, its Founder and Director, estimates “improved efficiency by 25–35%,” particularly in coding, testing, and documentation.
This trend is echoed in a report by SLASSCOM, the national chamber for Sri Lanka’s IT/BPM industry titled the Employability Skills Report 2024. The report compiled in collaboration with Deloitte, found that 51% of companies are already using GitHub Copilot, an AI-powered coding assistant, with another 56% planning to adopt it within the next 18 months. Across the 37 companies the report surveyed, similar tools such as AWS CodeWhisperer, Synk Code, and Codeium were also identified, reflecting the rapid uptake of AI-assisted developer tools and cloud frameworks across the industry.
Furthermore, GenAI is lowering barriers to software creation. Balathsan Sayanthan, a Founding Member of the Yarl IT Hub, a non-profit promoting technology and entrepreneurship in the Northern Province, cites their recent Build with AI for Everyone workshop in Jaffna, where 50 participants, most without technical backgrounds, built working web applications in just four days. “One participant was a teacher who had never written code. He built a full application, added AI features, and is now preparing to launch it as a startup. A month earlier, he knew nothing about building software,” Sayanthan recalls. “What impressed me is that instead of relying on others to translate his ideas, he was able to build it himself.
Yet the limits are clear. “Can AI do an integration with SAP and create an application to meet enterprise requirements? No. For that, you need humans,” Mangala notes. He cautions against assuming that AI is ready for full-scale deployment, arguing that its value lies in prototyping and support rather than enterprise-grade systems.
Taken together, these developments show that GenAI is already reshaping software development. It delivers efficiency gains within firms and opens opportunities for people without technical backgrounds. At the same time, its limitations are evident, particularly in scaling to complex tasks and ensuring long-term maintainability. Humans remain central to the process, but GenAI is expanding what’s possible while redefining the boundaries of expertise, placing greater value on problem-solving, domain knowledge, and sustained oversight.
How Is Generative AI Disrupting the Technology Job Market?
The sharpest impact of GenAI is visible at the entry level, where fresh graduates are struggling to find jobs. A key driver of this has been shifting client expectations owing to GenAI and the productivity gains it offers. “If you previously needed a team of ten developers, now clients ask for five who can use GenAI. They want to cut costs, so we can’t hire as we used to. We keep getting requests for internships, but we cannot take them,” says Mangala.
Global research confirms routine coding is among the first roles to be automated. The University of Stanford’s 2024 AI Index finds entry-level development tasks more vulnerable than higher-order work like systems architecture or cybersecurity. Sri Lankan industry leaders see the same pattern. The result is a shift in what companies value. Employers are no longer hiring graduates who can only write code, but those who can oversee AI outputs and apply domain knowledge.
“The premium skill is no longer programming syntax. Each customer, each vertical, has its own nuance,” Hariharan explains. Firms now prioritise candidates who combine AI literacy with domain expertise to solve real-world problems. For Sri Lanka, where GenAI is reshaping hiring expectations, the first rung of the career ladder is eroding. The roles that remain demand a higher bar of expertise, forcing many new graduates and even existing professionals to reskill if they are to remain in the industry.
Yet even as GenAI raises expectations, it hasn’t lifted pay for most. “The salaries used to increase annually, but now they’re stagnant,” says Mangala. GenAI-driven productivity has reduced the need for large developer teams, creating an oversupply of programmers, and slowing new hiring, which together have eroded the wage growth that defined the past decade. At the same time, a divide is emerging. Graduates who rely on AI only to generate code are easily commoditised and see little wage growth. At the other end, engineers who can conceptualise, fine-tune, and deploy AI systems are commanding premiums. “That pool is small,” notes Sayanthan, “but when they have such experience, they can demand a premium.” Those with these highly sought-after AI skills may even bypass local employers altogether, offering services directly to foreign clients while remaining in Sri Lanka.
“If you previously needed a team of ten developers, now clients ask for five who can use GenAI.”
The Rising Cost of Training New Talent
The disruption of entry-level pathways due to GenAI has also exposed long-standing weaknesses in how Sri Lanka trains new tech recruits. Fresh graduates often enter the industry with only basic skills, forcing firms to invest heavily in bringing them up to speed. “For six months it’s only training. We’re not making money on any of them,” says Mangala. He notes that while some hires can be productive within three months, “for most, it takes longer.” This comes at another cost: time senior engineers must spend supervising juniors instead of focusing on client work.
SLASSCOM’s Employability Skills Survey 2024 reflects this reality. Across the 37 firms surveyed, the report found that companies spend between Rs.100,000 and Rs.500,000 per graduate on training, with ramp-up periods of three to six months before new hires meaningfully contribute to projects. But with GenAI boosting productivity and clients pressing for leaner teams, firms are finding it increasingly difficult to make these investments. Fewer mentors and tighter budgets mean the traditional model of extended on-the-job learning is being tested just as AI reshapes what skills the industry now demands.
In response, some companies are turning to AI itself for help. At Cut+Dry, Shanil Fernando shares that they have deployed an AI agent trained on their codebase. “What used to take two months can now be learned in two weeks,” he says. “A fresh graduate coming in doesn’t need a senior engineer anymore. AI is the mentor.” Yet, he notes that the same productivity gains that make this possible have also frozen hiring, except for exceptional engineers. His advice to graduates: invest this time in mastering AI tools, because in the new landscape, that has become key to employability.
“We’ve stopped hiring. The main reason is significant productivity gains from AI-enabled tools.”
But while companies and professionals are racing to adapt, universities are struggling to keep pace with the speed of technological change. As AI reshapes the skills employers demand, Sri Lankan education still relies on outdated material. “We are still teaching things like CD-ROMs in schools, which no longer exist in the world,” observes Hariharan. Curricula have yet to adapt to cloud platforms, API-driven systems, or GenAI tools, widening the mismatch between what graduates learn and what companies now require.
The speed of AI’s advance only sharpens this gap. Anthropic CEO Dario Amodei predicts nearly all code could be AI-generated within the next year. Shanil is sceptical of the timeline but concedes that the pace of change is “crazy.” He warns that unless universities adapt continuously, even a four-year degree could risk obsolescence by graduation. “Adoption of AI should be a core subject in any industry,” he says, “because every job will carry an AI component. If you don’t adopt that skill, you’ll be obsolete. It will be like using a computer or the internet today.”
Amid these pressures, alternative pathways are showing promise. In Jaffna, the Yarl IT Hub runs Uki, a six-month outcome-based training programme that prepares graduates to join tech companies or launch their own ventures. Sayanthan notes that 90% of its most recent cohort, though lacking formal qualifications, has secured internships or started startups. “That was a surprise even to us,” he admits, in a market where most graduates struggle to land internships.
Together, these pressures show how GenAI is not only reshaping demand for jobs but also exposing weaknesses in how Sri Lanka produces talent. Traditional training is costly and slow, universities are falling behind, and even bootcamps require scarce mid-level mentors. While firms are turning to AI to fill the gap, the real need is for graduates with AI literacy. Models like Uki point to possible solutions. In a job market where GenAI has narrowed entry-level pathways and raised the bar for employability, adaptation, reskilling, and AI fluency will determine who thrives.
The Missing Bridge Between Experience and Change
Software delivery rests on a clear hierarchy. Senior architects and product managers set direction and define system architecture. At the entry level, juniors are still learning to apply classroom knowledge to real projects. Between them sits the middle layer: senior software engineers, delivery managers, and team leads with five to eight years of experience. They break plans into tasks, mentor juniors, review code, and carry the context that turns strategy into execution. As Hariharan explains, “These are the people who deliver the work and train others to deliver it. When that layer goes, it slows everything down.”
■ 83.7% find it difficult to hire new employees with the amount of experience they require.
■ 95% find it challenging to find new employees with the right technical skills.
■ 81% agreed that there is an increase in competition when retaining or hiring highly skilled employees.
■ 38% had to hire new employees without the skills needed.
Sri Lanka’s AI moment arrived on top of a deeper wound. The economic crisis triggered an exodus of mid-level engineers who formed the backbone of delivery teams. The scale of the gap is visible in SLASSCOM’s 2024 Employability Skills Survey, which found that 84% of firms struggled to recruit candidates with the required experience. Migration ranked among the top three retention challenges. Hybrid work, now the norm, further slowed the transfer of tacit knowledge. “The knowledge is there. What’s missing is depth. That will only come with time,” says Hariharan, referring to the practical experience and contextual judgement that help firms navigate shifts like the current rise of GenAI. Without that depth, firms are left with fewer people able to bridge technical change with delivery reality.
The effects were not uniform across the island. In Colombo, where most large IT firms operate, the loss of mid-level engineers hollowed out delivery capacity. In the Northern Province, home to more than 150 small and medium-sized tech companies, the challenge was more fundamental. Years of brain drain had already weakened the talent base, while the COVID-19 pandemic saw many graduates leave before entering the local industry. “It wasn’t just experienced professionals. Instead of joining the industry, graduates were looking at foreign opportunities. The crisis was a trigger for them to leave,” says Sayanthan. As a result, Northern tech firms entered the age of GenAI with a thinner foundation of skills, making it harder to adapt to the pace of technological change.
Signs of recovery, however, are emerging. Shanil observes that outward migration has slowed, with more Sri Lankan professionals choosing to stay—helped by opportunities to work remotely for global firms while remaining in the country. At the same time, employers like his are no longer limited to local talent, as glo- balisation has expanded access to worldwide pools of expertise. This recovery is taking shape even as GenAI continues to redefine workforce structures and reshape how teams are built. “This disruption will continue for a time,” Shanil says, “but eventually it will stabilise, and what seems disruptive today will become the new normal.”
“It wasn’t just experienced professionals. Instead of joining the industry, graduates were looking at foreign opportunities. The crisis was a trigger for them to leave.”
Sri Lanka’s IT Industry at a Crossroads
Sri Lanka’s IT industry is navigating a rare convergence of global and local shocks. GenAI has automated routine coding and narrowed the path for entry-level jobs just as the country is still recovering from the loss of its mid-level talent during the economic crisis. Training costs remain high, universities are struggling to keep pace, and firms are demanding leaner, AI-augmented teams. For graduates, the first rung of the career ladder has shifted, demanding domain expertise, adaptability, and AI fluency as baseline skills.
Yet, the industry doesn’t see GenAI as a threat but a catalyst. Shanil Fernando argues that it can lift productivity “to levels never seen before,” provided the workforce embraces AI. Mangala agrees the current hype cycle will stabilise, but stresses the need to grow more companies to absorb the country’s expanding graduate output. And Hariharan Padmanaban points to the wider multiplier effect of IT jobs, where every tech hire sustains four others across food, transport, and services.
The consensus is clear: GenAI is an opportunity. The test is whether Sri Lanka can adapt quickly enough: reshaping training, modernising curricula, and building an ecosystem that allows AI-enabled engineers and startups alike to thrive.