That era is officially ending. A tectonic shift is ripping through the billing departments at Infosys, TCS, and Wipro as “agentic” artificial intelligence, or AI systems capable of independent work in real-time sensemaking situations, becomes commonplace. Unlike ages-old chatbots, which may be able to pen you a poem or sum up an email, however, Agentic AI — which is being driven by platforms including the most recent entry, Claude Cowork — doesn’t just chatter on; it gets things done. It runs complex workflows, debugs code, and migrates servers so well that the traditional role of a junior engineer is no longer needed. Results-oriented change that won’t convert to compliance charging. As a result, you will find that clients are not interested in paying for ‘effort.’ They want to pay for “results,’’ and it’s forcing an ugly, necessary migration in how companies buy and sell software services.
Defining Agentic AI
To understand the disruption, let’s differentiate ‘Generative AI’ from ‘Agentic AI’. Generative AI is akin to the really smart librarian who can read every book and write a report for you. Agentic AI is like hiring a contractor who can physically build the house. Tools like Claude Cowork can be provided with a high-level objective, ‘update the payment gateway for the legacy app, and the AI fills in the gap – discovering prior art, rummaging through DB code to find what has already been done, coding it up, testing, and deploying, and only asking humans when it gets stuck. In the classic approach, you might have needed a team of five developers, budgeting 40 hours per developer. If an AI agent can do the job for a tiny fraction of the expense in four hours, the clients will guffaw when you present them with a bill based on ‘headcount’.” The seat model, which holds that revenue equals the number of bodies thrown at a problem, is disappearing.
The New Currency: Productivity
This change is prompting a scramble over new metrics. Specialists say there’s a shift toward “productivity-linked” contracts. IT firms are not just about billable hours anymore; they have to deliver fixed-price outcomes. If the AI can build the team in half the time, the client wants a 50% price cut or, at a minimum, have that value substantially compressed into their contract. It’s a two-edged sword for Indian IT behemoths. On the one hand, they can use these AI agents to do more work with fewer people and may be able to raise their profit margins. On the other hand, their top-line revenue — which has historically been bloated by army upon army of entry-level coders — is under attack. They are in the business of peddling efficiency, not labor. The IT ‘pyramid model’ that relied heavily on a base of young graduates is turning upside down.
The Human Impact
What does this mean for workers? It is an existential pivot. The ‘middleman’ jobs — repetitive testing, basic maintenance, and data entry — are being gobbled up by brokers. A human worker is now valuable only as an orchestrator or supervisor. The engineer of the future is not a coder but an AI-robot whisperer. They are the maestros of a digital symphony. This transition is messy. We have contract renegotiations, a pullback on hiring freshers, and a mad rush to upskill employees to manage these new agentic platforms. The industry is transitioning from a factory-floor model to one based on high-end consultancy, and not everyone will make the transition.
