The GCC Is Where Transformation Happens — Not Where It Gets Outsourced

I built one of these before the category had a name. At GE Capital in Gurgaon, starting in the late 1990s, we stood up analytics and operations centers from the ground up — the work that the industry would later label the Global Capability Center. At the time there was no playbook and no acronym. There was a floor, a hiring problem, a set of processes that ran badly in New York and Stamford, and a thesis that they could run better somewhere else if we rebuilt them rather than relocated them.

That distinction — rebuilt, not relocated — is the whole story, and it is the part the current panic about AI and offshoring keeps missing.

What the GCC Was Actually For

The lazy history says the GCC was a cost-arbitrage machine: move the work where the labor is cheaper, bank the difference. That is the version on the slide, and it was never the version that created durable value. The centers that mattered did something harder. They took processes that had grown up by accretion at headquarters — undocumented, owner-dependent, riddled with exceptions nobody could explain — and they were forced to write them down, standardize them, and run them at volume with people who had not inherited the institutional folklore. The act of moving the work made you understand the work. That understanding, not the wage differential, was the asset.

Two decades on, the typical financial-services GCC is no longer a back office. It owns analytics, model development, large parts of technology, and increasingly the decision logic of the institution itself. It is where the process documentation actually exists. It is where the operating discipline lives. And it is staffed, at this point, with some of the densest concentrations of quantitative and engineering talent the institution has anywhere.

Why the Cost Case Is Collapsing

Here is the uncomfortable part. The work the GCC was originally built to do cheaply at distance is exactly the work AI now does outright. High-volume analytical processing, reconciliation, first-pass model runs, document handling — the categories that justified the arbitrage are the categories automation absorbs first. So the original business case is genuinely eroding. An executive looking only at the cost line is right that something is changing underneath it.

What that executive gets wrong is the conclusion. Reading the GCC as a cost line, the move looks obvious: AI does the cheap work now, so shrink the center that did the cheap work. That is the same fingers-and-toes error that defines most failed AI programs, just exported to a different continent. You automate the task, you cut the people who did the task, and you congratulate yourself on a saving — while quietly disposing of the one organization in the enterprise that knows how the process actually works and has the discipline to redesign it.

Where Transformation Actually Lands

Transformation does not happen on the slide at headquarters. It happens where the work lives. And for most financial institutions, the work has lived in the GCC for twenty years. The center is where the processes are documented, where the exceptions are understood, where the talent dense enough to rebuild the operating model around AI is already sitting. If you are serious about taking the old way of working apart so the new way has somewhere to live — the only version of AI that ever reaches the P&L — the GCC is not the thing you cut. It is the thing you put in charge of the cutting.

The institutions that understand this are quietly doing the opposite of the panic move. They are elevating the GCC from a delivery center to a transformation center — giving it ownership of the redesign, not just the execution. The captive center stops being where work gets sent and becomes where the operating model gets rebuilt. That is a far better use of two decades of accumulated process knowledge than treating it as a headcount number to optimize.

The GCC was never really about doing the work cheaply. It was about being the one place forced to understand the work well enough to run it deliberately. That capability is worth more now, not less. The institutions reading their center as a cost to remove will spend the next decade rebuilding, at a premium, the operating discipline they already owned. The ones reading it as where transformation happens will have a twenty-year head start on the only work that matters.