The Two Layers of AI
To understand this shift, it helps to think of the AI world in two main layers. First, there's the 'model layer'. This is where foundational models like GPT-4 are built. It's a high-stakes game played by a handful of global tech giants, requiring billions
in investment and massive computing power. Sandip Agarwal, a fund manager at Sowilo Investment Managers and a respected voice in India's tech sector, suggests that the most intense phase of this hardware-driven boom is maturing. The second, and increasingly important, layer is the 'application layer'. This is where the magic happens for most businesses. It involves taking the raw power of a foundational model and building a specific, useful product around it—like a customer service chatbot, a legal contract analyzer, or a medical transcription tool. This is where Agarwal sees the opportunity broadening significantly.
From Building Engines to Building Cars
A helpful analogy is the automotive industry. The model layer is like designing and manufacturing a powerful V8 engine—a complex, capital-intensive process. The application layer is everything else: building the car's chassis, designing the interior, adding the wheels, and creating a vehicle that someone would actually want to drive. You can't drive a raw engine to the grocery store, and similarly, most businesses don't interact directly with a raw AI model. They use applications. This is the core of Agarwal's point: while the world has been mesmerized by the engine builders, the real growth is now in the car factories—the companies that are putting these powerful AI engines to work to solve specific problems.
Why This Shift Matters for India
This evolution is particularly crucial for India. Competing at the foundational model layer requires a level of capital and infrastructure that is hard to match. However, India's tech industry is perfectly positioned to dominate the application layer. With its vast pool of skilled engineers, a thriving IT services sector, and deep domain expertise across various industries, India can become the world's 'software factory' for AI-powered solutions. According to Agarwal, while AI may reduce the effort required for some traditional IT tasks, it opens up a massive new opportunity in servicing enterprise AI needs. Indian IT firms are already retraining their workforce at a massive scale, preparing them to meet this new demand. Recent reports show a surge in domestic AI investments, with 80% of Indian organizations investing in technology to drive revenue growth.
Where Are the Opportunities?
The opportunities in the AI application layer are vast and cut across every sector. In healthcare, AI is being used for diagnostics and to help create personalised treatment plans. In finance, it's powering fraud detection and algorithmic trading. E-commerce platforms use it for hyper-personalization, while manufacturing plants deploy AI for predictive maintenance and quality control. The Indian government's IndiaAI Mission, backed by over ₹10,000 crore, is further accelerating the development of India-specific applications in agriculture, healthcare, and governance. This move towards 'sovereign AI'—using domestic data and infrastructure—is creating a unique and protected market for innovation.
The Rise of the AI-Enabled Professional
This shift also changes what it means to work with technology. The most valuable skill is no longer just being able to code an AI model from scratch. Instead, there is a growing demand for 'AI-enabled' professionals. These are doctors, lawyers, marketers, and product managers who understand how to use AI tools to make their work more effective. Careers are emerging for AI Product Managers, who bridge the gap between technical teams and business needs, and Natural Language Processing (NLP) Engineers, who create systems like chatbots. The future belongs to those who can creatively apply AI to their domain, not just those who build the underlying technology. It signals a move from AI being a niche for specialists to becoming a baseline professional skill.
