AI didn’t shrink the GCC’s job. It redefined who owns the decision

Walk onto a GCC engineering floor today and you will see something that would have been unthinkable three years ago. Senior engineers are spending more time reviewing an AI agent’s code than writing their own, because the agent already produces working software in minutes. What they are really reviewing is not the code. It is whether they trust it enough to ship it, and whether anyone could explain why it failed if it does. Writing code is being replaced by auditing judgment, and that shift is a preview of what is about to happen to every function inside the enterprise, not just engineering.

Ask any enterprise how far along they are on AI, and the answer is usually a number: copilots deployed, percentage of code AI-generated, models in production. That number is the wrong scoreboard.

The real question enterprises face now is what happens once AI is no longer a tool layered on top of work, but the foundation work runs on. That question does not get answered by a deployment metric. It gets answered by how an organisation reprices its own economics: which activities it automates, which capabilities it pays a premium for, and where accountability sits once machines begin doing more of the execution.

India’s GCCs are where enterprises are answering that question first, because GCCs now carry a mandate spanning engineering, product, AI, data, cybersecurity, and transformation. That gives GCCs a vantage point on the AI-era enterprise that few other functions have. Three repricings are underway at once: of work, of workers, and of the workplace itself.

Work Is Shifting From Volume to Judgment

AI is changing the economics of the work itself—routine execution is being automated at pace, while work demanding technical depth, domain judgment, and governance is commanding a premium.

Zinnov’s Salary Increase, Attrition and Hiring Trends 2026 report projects GCC hiring growth of 27.4% in 2026, with ER&D hiring running higher at 30.8%. That growth does not mean every role or layer of the workforce expands with it. AI will eliminate tasks faster than entire occupations, and the same business output may require fewer people at one layer while creating shortages of more expensive talent at another. Some roles will shrink, some will split into higher- and lower-value work, and new ones will emerge. The more consequential signal is where demand is concentrating: towards AI engineering, prompt engineering, forward-deployed engineering, AI forensics, and data governance, even as routine-heavy roles begin to contract. That shift is also visible in pay. Against an average GCC salary increase of 9.8%, AI/ML talent is being repriced at 21.1%.

Engineering leaders are increasingly pointing to the same underlying shift: as AI drives the marginal cost of producing code towards zero, the real premium is moving to deciding what should be built in the first place. The differentiator is no longer simply the ability to write code, but the judgment to identify the right problems, make the right trade-offs, and determine what is worth building.

The same economics will travel into other functions. An underwriting engine may assemble a risk view in seconds; a security agent may triage thousands of alerts before an analyst opens a dashboard. But the consequential decision—which risk to take, which signal to escalate, which output to trust—still needs an owner. Once execution becomes abundant, judgment becomes scarce.

Talent Is Moving From Tenure to Capability

Experience and credentials still matter, but they are weakening as proxies for future value as the half-life of skills shrinks. Our research shows this playing out as skill hopping and degree-blind hiring: a market weighting demonstrated capability over pedigree. As AI tools become easier to access, fluency in any one tool is unlikely to remain a durable differentiator. The premium will sit with people who can combine AI fluency with domain depth, business context, and judgment.

Reward structures are following. High performers are projected to see salary increases of 17.2% in 2026, against 9.8% overall. For GCCs, the resulting advantage will come less from the depth of the talent pool and more from the concentration of AI expertise, domain depth, and applied AI fluency within the same teams.

There is also a deeper risk hiding inside the productivity story. Routine work has always done more than produce output; it has been the apprenticeship layer through which junior employees learnt the business, made low-stakes mistakes, recognized patterns, and eventually developed judgment. AI is beginning to remove that work before it removes the need for the judgment it used to create.

If AI eliminates the apprenticeship work, where will the next generation of experts come from?

Enterprises can no longer assume judgment will emerge naturally through tenure. It will have to be built deliberately—through exposure to real decisions, domain rotations, supervised ownership, and progressively harder problem sets. A GCC that automates the base of its pyramid without redesigning how the next layer of decision-makers gets built will be capability-rich today and leadership-poor in five years.

The Workplace Is Moving From Fixed Roles to Continuous Adaptation

When work and workers are in continuous motion, the workplace cannot remain built around fixed roles and static teams.

The traditional enterprise was designed around clear job descriptions, functional silos, annual planning cycles, periodic training, and decisions flowing through hierarchy. AI is beginning to challenge each of those assumptions. Teams can be assembled around outcomes rather than functions. People can move across roles as capabilities change. Learning can sit inside the flow of work rather than outside it. Managers will spend less time allocating tasks and more time building judgment, directing experimentation, and deciding where human attention creates the most value.

The bigger change is in how capacity itself gets allocated. As AI makes capacity more elastic, annual headcount planning and fixed team structures become increasingly blunt instruments.

Capacity released through AI cannot simply be absorbed back into the same organisation. It has to be redirected towards faster decisions, deeper customer work, new products, or mandates that were previously uneconomic. Otherwise, productivity improves without materially changing enterprise value.

Performance will have to move as well—from measuring activity and headcount towards measuring value created, capacity released, and higher-order work enabled. This is where GCCs hold a structural advantage. Their proximity to enterprise mandates, combined with deep digital talent ecosystems, gives them room to experiment with team structures, role design, learning models, and AI-enabled workflows at a pace that is harder to achieve across the broader enterprise. The strongest GCCs can become places where new ways of working are tested, proven, and then taken global.

The question for GCCs is therefore no longer simply how much work can be moved to India. It is what parts of the enterprise can be redesigned through India.

The GCC that wins in the AI era will be the one that converts those ingredients into something tangible: higher-value work, stronger judgment, faster decisions, and mandates that carry real enterprise accountability.

(The author is President, Globalization Excellence, Zinnov. Views are personal.)