Back in the early 2000s, Indian IT workers engaged in low-end, repetitive, and routine tasks – especially in call centres and BPOs – were often described as cyber coolies. Is this infamous term raising its head in the age of AI too? As bot coolies? Despite a large and growing pool of engineers and tech talent, the tech industry has lagged in creating foundational models. Many IT companies are more focused on applied/custom AI services than cutting-edge foundational research. A lot of their work is about building and deploying AI solutions for clients, rather than doing pure research.
The greater concern is that India might simply replay the low-margin IT services model in a new AI guise. “It is true that compared to the US and China, we have been late innovators in AI but our future will lie in applications and adaptation which has long been our forte,” says Ganesh Natarajan chairman, GTT Data Solutions & 5F World. “Our decades of business and application skills must be brought to think through new end-to-end processes for organisations, train consultants in prompt engineering and GenAI and technologists to design and deploy AI agents across the reengineered applications value chain,” he adds.
Many large Indian IT companies are fundamentally service providers. Their business model has historically been about outsourcing, customisation, and implementation.
What did Soumyakanti Chakraborty say?
According to Soumyakanti Chakraborty, professor, management information systems, IIM Calcutta, most AI work today focuses on integrating, customising, and building applications on top of foreign foundation models. Global companies are expanding their delivery centres here and leveraging Indian talent for AI services, but rarely for core model research and development.
“However, it would be inaccurate to say no deeper work is happening,” says Chakraborty. There are notable efforts to create homegrown large language models and language technologies – such as Sarvam AI, Krutrim, AI4Bharat,OpenHathi etc. The IndiaAI Mission supports such foundational models with funding, compute, and collaborative frameworks. “While the risk of India being relegated to a “bot coolie” role exists, it is not our destiny yet. The AI ecosystem is still evolving, and India’s position remains flexible,” he adds.
“The fear of obsolescence should only be for dinosaurs who refuse to adapt to changing environments. The smart folks will lead the reimagining of processes and systems and play pivotal future roles,” Natarajan says, adding, “the new era demands dual intelligence with humans understanding the potential of AI agents and deploying them in new work processes.”
India is comfortable with the IT services model, but fundamental AI research requires patient, long-term investment and higher risk tolerance. “Many of India’s best AI researchers move abroad or join global tech giants, while domestic IT firms often lack world-class research facilities and treat AI research as a cost centre rather than an investment. To break free from the “bot coolie” narrative, Chakraborty says that Indian firms must treat AI as a business priority with clear multi-year investment horizons, supported by long-term partnerships and joint research with premier academics.
About foundational AI
Foundational AI is built by deeptech innovators, not by services firms optimised for delivery efficiency, predictability, and global compliance, says Harish Mehta, founder & executive chairman of Onward Technologies. “Asking our IT giants to create world-leading AI is like asking an airline to manufacture aircraft – because while both use “software,” their DNA, incentives, and success metrics are different. We need a two-engine strategy: services firms driving large-scale AI adoption, and scientist-founders backed by risk capital building original models, chips, and IP,” he adds.
The real concern lies in the development of India’s AI sector. “Unlike the BPO era where we controlled service delivery infrastructure, today, we’re building AI innovations on foundations we don’t own,” says Deepa Mani, professor of information systems and deputy dean, Academic Programmes, Indian School of Business. “Our research shows 67-83% of Indian AI companies operate exclusively in the application layer – implementing solutions built on foundational models and infrastructure controlled by Google, Meta, and other global firms. The quantum and growth in startups working in the foundational layers in the US is substantively greater than that in India, exposing the risk that we are sophisticated users rather than architects of the technology,” she adds.
Also, India’s challenge isn’t talent – it’s a system that fails innovators, according to Mehta. “While the US and China fund foundational models, our entrepreneurs are still chasing signatures and approvals. We must simultaneously reform governance while sharply increasing R&D investment across both public and private sectors to truly become an innovation-driven nation,” he emphasises.
Mani feels that India needs to recognise key challenges like capital constraints, data asymmetry, weak research ecosystem, governance gaps, and swiftly address them. AI is a general-purpose technology that has systematic and pervasive impact on value creation across a swathe of sectors, organisations and functions. “India must invest in sovereign AI capabilities else risk dependent status in one of the most fundamental technologies of our time,” she adds.
