GCCs are de-coupling from the parent, and hiring hasn’t caught up

The most consequential shift in BFSI GCCs is the quiet moment the parent stops handing down finished decisions and starts handing over the decision itself. The centre goes from executing a charter to owning it. And the leader who was excellent at running the first is often the wrong hire for the second.

Running a charter rewards throughput, predictability, and a clean handshake with the parent. Owning one rewards judgment under ambiguity, the standing to say no to the parent, and the appetite to be accountable for an outcome nobody upstream is going to underwrite for you. Those are different people. The market has far fewer of the second kind, and they’re rarely the ones already sitting in a delivery-leadership seat, waiting to be promoted into it.

De-coupling shows up in the language a centre uses about itself long before it shows up in a job spec. The parent starts describing the centre as a business it holds rather than a vendor it manages. Roadmap ownership moves. So does the authority to kill a project the parent was fond of. By the time an outside observer would call the centre de-coupled, it’s usually been making its own consequential calls for a year, and hiring for a version of itself that no longer exists.

The first thing to move is authority, ahead of the org chart. When a centre owned nothing, the parent effectively approved its senior hires, a regional HR leader or a functional head upstream had the final word, and the centre’s role was to shortlist and defend. When a centre owns its charter, that arrangement quietly inverts. The person accountable for the outcome has to be the person who signs the leader responsible for delivering it.

Most centres haven’t made that shift cleanly. The title on the requisition says the centre owns the hire; the actual sign-off still routes through the parent, through a compensation committee calibrated to a market the parent understands and the centre has outgrown. The result is a hire the centre is accountable for but didn’t fully control, approved against the parent’s idea of the role, which no longer matches the centre’s reality. That’s a governance problem dressed as a hiring one, and it surfaces first at the offer stage, where the parent’s band and the centre’s market collide.

Alongside who signs, the profile of the person being signed moves too. A delivery leader is hired for reliability against a defined scope. An owner is hired for judgment against a scope that’s still forming. On paper the two roles can read almost identically: same seniority, same function, same reporting line. Underneath, they’re different labour markets, priced and sourced differently, the way a skill and its proof are different things.

Here’s the gap we keep seeing. The centre has already de-coupled in practice, it’s making calls the parent used to make, but the job specs still read like delivery-leadership roles. Same competencies, same “partners with the parent to execute” language, same interview loop built to test whether someone can run a known process well. The org grew up; the hiring didn’t. So the centre screens hard for the wrong thing and screens out the people it actually needs, because an owner interviewing against a delivery spec looks over-qualified and under-controllable at the same time.

Closing that gap means searching differently. The candidates who can own a charter rarely announce themselves as available, and they don’t screen well against a delivery-shaped spec. In BFSI especially, where the risk and regulatory stakes make ownership real, this is the hire that decides whether de-coupling is a promotion for the centre or a title change that nothing backs.

The same lag that leaves the spec behind the reality leaves the compensation behind it too. A de-coupled centre competes for a narrower, scarcer kind of leader, and for the specialist seats underneath that leader, in specialised AI, quant, and applied data inside BFSI, the market can move meaningfully in a single quarter. The annual comp band is a fine instrument for a stable market. For those seats it’s stale before the ink dries: a band finalised in the annual cycle already lags by the time the first offer goes out against it, and the gap only widens as the year runs.

The centres feeling this most acutely aren’t losing candidates on the average role. They’re losing them on the sharp end, the handful of seats where the market repriced between the band being set and the offer being made. Those are precisely the seats a de-coupled centre can least afford to lose, because they’re the ones the charter now depends on. And the published band was never the real number anyway; the published band and the closed offer are two different numbers, and the distance between them is widest exactly where supply is thinnest.

The smarter BFSI centres are splitting the clock. Most roles stay on the annual cadence, where predictability and governance matter more than speed. The fast-repricing seats get moved onto a live view of the market, refreshed as the market actually moves, so the offer reflects this quarter rather than last spring. The discipline is knowing which seats belong on which clock. Put everything on the fast clock and you’ve traded governance for chaos; keep everything on the slow one and you keep bleeding the scarce hires. In a regulated industry, the art is running both cadences without letting the fast one contaminate the bands that are supposed to hold.

The lag is structural. A centre’s operating reality changes through hundreds of small decisions, a call made here, a project owned there, none of which triggers a rewrite of the leadership spec or the comp philosophy. Those documents change on a governance cadence: a review, a committee, an annual cycle. So the operating model runs ahead on a fast clock while the hiring machinery follows on a slow one, and the two only reconcile when a search goes badly enough to force the question.

The parent rarely sees it either. It reads a single leadership band and a single market average and assumes they still describe the centre. They no longer do. The parent’s frame was built when the centre executed; the centre now owns, and owning carries a different price and a different person. Both cost more, and both are harder to find. The averages the parent is working from are true about the old centre and useless about the new one.

This is why de-coupling is a leadership problem before it’s a mandate problem. The mandate can be granted in a memo. The leadership to carry it, and the pay to secure that leadership, both have to be recruited into a market that priced them the moment the charter changed, and the hiring bar has to move first, or the mandate arrives with no one able to hold it.

Start with sign-off. If the centre owns the charter, the centre leader accountable for it needs the final word on the leaders who deliver it, with the parent moving from an approval role to a governance one. Until that authority actually moves, everything downstream, the spec, the search, the offer, is calibrated to the wrong party, and the hire the centre is accountable for stays outside its control.

Then rewrite the spec to the reality. Describe the job the centre actually now does, the decisions it makes, the outcomes it owns, the calls it takes without the parent, and interview for judgment under ambiguity. The candidates who fit that description don’t come from the delivery pipeline, so the search has to reach into a different pool and often approach people who aren’t looking.

Last, put the scarce seats on the right clock and pay them against the market that actually prices them. For those roles, the published band is a starting reference, not the offer, and the centre has to be ready to close case by case against what the market did this quarter. Get sign-off, spec, and clock aligned to the charter the centre already owns, and de-coupling becomes real. Leave them calibrated to the centre’s old life, and the title changes while the capability doesn’t.