Artificial Intelligence (AI) has moved from experimentation to enterprise adoption in under two years, with industry estimates projecting AI services revenues of $10 billion-$12 billion in FY26. Yet, this moment of expansion coincides with layoffs, automation, and the vulnerability of entry-level roles in India’s Information Technology (IT) and Business Process Outsourcing (BPO) sectors. It also coincides with India’s top tech firms integrating AI products into their workflows. Have AI products/LLMs started to disrupt the software services industry? Kishan Sundar and Alaganambi Welkin discuss the question in a conversation moderated by Kunal Shankar. Edited excerpts:
Are the rapid advances in AI disrupting India’s IT services sector, or are we witnessing a transformation?
Alagunambi Welkin: Based on discussions with union members across companies and levels, what we think is happening now is AI washing. The retrenchments or restructuring happening in companies are not primarily due to AI, but in the name of AI.
At the same time, transformation is happening. Every developer and key engineer in IT services companies is adapting to AI tools that assist their day-to-day life. There is a wide shift happening across the industry in how we work, and that transformation is positive.
But the claims that AI is going to take over massive jobs, that no more developers are needed, and that entire end-to-end development can be done by AI… we think that is far-fetched. So, what we see is cost cutting, typically done by major multinational corporations, now being termed AI-based cost-cutting.
Kishan Sundar: AI is not replacing the industry; it is transforming it. In the last three decades, the industry has operated on labour arbitrage. Growth came from adding people. That was its strength — well-managed processes and predictable delivery. Now we are moving towards intelligence arbitrage. GenAI enables growth without a matching increase in staff.
From COBOL days, when you had to remember syntax and semantics, to IntelliSense improving productivity, to AI assistance that completes methods and refactors — today, by giving a prompt, you get well-structured code. But is that code production-ready? You need a different skill set to assess whether it is good enough, whether it needs refactoring, whether the right pattern is adopted.
The same shift is happening for product managers. Writing user stories and acceptance criteria, which used to take hours and required concurrence from developers and testers, can now be done much faster. Testers are becoming efficient. DevOps engineers are becoming efficient. In essence, a squad of eight to 10 members can become three to five.
But does that mean roles are shrinking and layoffs are inevitable? No. AI applications are not deploy-and-forget. There is drift and fine-tuning. Traditional AMS (application management services) now need to adopt differently. Additional roles are getting created.
Net-net, revenue per engineer is increasing. The number of people needed per engagement may reduce, but the range of roles is expanding.
Are entry-level roles, especially in BPO and KPO (knowledge process outsourcing), more vulnerable?
AW: When we speak about the IT industry in India, there are two major classifications. One is IT services companies that build, maintain, and enhance software. The other is BPO and KPO jobs, which are repetitive and well-defined.
Because of the scale, we increase head count and deliver through documentation or calls. Those kinds of jobs are definitely vulnerable now. With agentic AI developments, companies are saying they can automate end-to-end processes. A call centre with 4,000-5,000 staff does not need all of them for every process. Handling validation and retraining in response to deviations may require just 10–15 people.
On the other side, in IT services companies, AI is assisting the end-to-end software development cycle and drastically reducing the number of hours required to complete a task.
But development is not just coding. Engineers interact with multiple teams — within the client organisation, across geographies, and with counterpart teams in the U.S. or the U.K. There are interdependencies, internal politics, differing agendas, and competition. Even if software can be built in two or three hours, it still requires requirements from one team and data from another. And when someone feels vulnerable, they may not communicate those requirements clearly. So, a part of the job is getting easier because of AI tools, but that alone is not enough to replace humans entirely. At the same time, if the hours required to build software drop significantly, there may be a case for reducing overall working hours.
Are global AI partnerships defensive moves, or growth strategies?
KS: They are definitely growth strategies. We should not fall into the trap of vibe coding. In regulated environments such as banking and financial services, every line written must have audit and traceability — why it was written and for what purpose. We are adopting the same Software Development Life Cycle but embedding AI in every aspect. LLMs generate code. We build wrappers and context layers so that every line generated against a prompt is repeatable and consistent across developers. That ensures maintainability.
Gone are the days of the 10x developer. Now context engineering is the essence. The engineer who knows the domain and context becomes critical rather than someone who codes fast.
Most LLMs are horizontal. Where we come in is bringing domain context — retail banking, wealth management, corporate banking — and embedding regulatory aspects into the lifecycle.
Beyond services, we are developing proprietary frameworks and product engineering capabilities.
Is India largely consuming AI built elsewhere?
AW: The MNCs leading LLMs and AI transformations own the IP and build foundational models with enormous capital and infrastructure. Indian IT companies collaborate with them, use their services, rebrand them, or build on top of them. Compared to the U.S. or China, we invest insufficiently in education, research, compute capacity, and data infrastructure, so we are not building foundational models that can compete globally. Reskilling is happening mainly on the consumer side — prompt engineering, context engineering, and agent building.
KS: It is not completely black and white. India has strengths in systems engineering, enterprise integration, scaling, execution, and process rigour. The strategic question is whether we prioritise sovereign LLMs or double down on AI services dominance. Both should happen, but prioritisation matters.
Is the services model shifting from manpower-driven billing to outcome-based pricing?
KS: We moved from the traditional pyramid model to a diamond structure and from time-and-material pricing to squad-based pricing. Now, we are moving further towards outcome-based or output-based pricing. Customers care about predictable delivery, quality, and clarity of cost upfront. With outcome-based pricing, what matters is predictable delivery and quality. That gives us the opportunity to grow faster while maintaining margins.
Has the time come to speak about a ‘just transition’ for the IT sector? What should companies and government do, and what protections are needed?
AW: That part is not being discussed enough. When transformation happens, unemployment increases. If a person suddenly gets laid off without information, his financial planning, family, and children’s education are all affected. Employees in India pay huge taxes. Why don’t we get unemployment benefits when we do not get jobs for six months or a year?
Skill India is not credit-based. You cannot formally certify your skills in India through it. Simply watching a video and claiming competence is not enough for the industry.
Algorithms are increasingly deciding work and life. There should be transparency and regulation. Otherwise, it will create difficult situations, including mental health concerns.
Data centres are going to increase because of AI and data protection needs. They create less employment but more climate impact — electricity consumption and water usage. That impact also needs to be discussed.
About the Author
As the Chief Technology Officer, Kishan Sundar helms the technology strategy for Maveric. His leadership in creating engagement and impact through customized technology solutions and emerging technologies will play a crucial role in accelerating Maveric’s revenue growth and fuelling its aspiration of becoming one of the top three Bank Tech companies.
