The Salaried Worker vs. The AI Task
For decades, the cost of a knowledge worker has been a predictable, bundled package. A company pays a salary and in return gets an employee's time, expertise, creativity, problem-solving skills, and ability to perform a wide range of tasks. The fully
loaded cost includes not just the salary, but benefits, training, office space, and management overhead. This model is predictable but also inflexible; you pay the full cost whether the employee is running at 110% capacity or having a quiet week. AI introduces a completely different economic model: the cost per task. Instead of buying a person's time, a company can now buy a specific output. Whether it's generating a paragraph of text, analysing a dataset, or responding to a customer email, AI services are often priced on a usage basis, like paying for electricity.
A Look at AI's Price Tag
AI pricing isn't a single, simple number. For businesses using foundational models like those from OpenAI or Anthropic, costs are often measured in 'tokens'—small pieces of words. A company might pay a fraction of a rupee for an AI to process an email, scaling linearly with usage. For example, generating a business email might cost less than a few rupees in token fees. Other models use a per-seat subscription, similar to traditional software, with enterprise plans often costing between $20 to $60 per user per month. This unbundling of work from a single salary creates a granular, variable cost structure that can be incredibly efficient for high-volume, repeatable tasks. At a large enough scale, the cost per task for an AI can be exponentially cheaper than a human equivalent.
The Hidden Costs of 'Cheaper' AI
This is where the headline's 'complication' comes in. While a single AI task may be cheap, the initial software subscription is often the smallest part of the total bill. Companies face significant hidden costs in data preparation, integrating the AI with existing systems, employee training, and ongoing maintenance. These operational expenses can add 35-50% or more to the original budget. Furthermore, there's the issue of what AI can't do. It struggles with nuanced judgement, complex interpersonal communication, and open-ended creative problem-solving. A salaried worker provides value far beyond their easily automated tasks, including mentoring, handling exceptions, and contributing to company culture—things that don't have a line item on a budget sheet. This is why some firms that rushed to replace staff with AI have found themselves quietly rehiring.
Implications for the Indian Workforce
For India, with its vast salaried workforce and world-leading IT and business process outsourcing sectors, this economic shift is critical. The debate isn't just about job losses, but a fundamental reshaping of roles. Studies suggest that while some jobs will be substituted, the majority of roles will be changed, with AI augmenting human workers and boosting productivity. A Goldman Sachs report noted nearly half of India's non-agricultural workforce could see productivity gains from AI. Already, over 40% of salaried Indians report that AI has improved their income or productivity. The challenge for Indian businesses and professionals will be to adapt. The future belongs not to those who can simply perform a task, but to those who can manage, orchestrate, and creatively apply these powerful new tools. It's a shift from being a salaried task-doer to a strategic problem-solver.
