The Conversation: 'My parent's CFO has a spreadsheet that doesn't have my name on it'

This week's Conversation is with Priya Subramaniam (a composite, drawn from interviews with six current Global Capability Centre heads in Bengaluru, Hyderabad, and Pune; details are de-identified). She has run a 5,200-person banking captive for a Tier 1 European parent for the past seven years. We spoke on Tuesday, two days after Google Cloud Next 2026 and a week after Cognizant's Project Leap. She asked that the conversation be unattributed; everything else is on the record.

---

Q1: Walk me through what changed for you the day Google announced the Gemini Enterprise Agent Platform.

What changed is not the technology. The technology has been visible for eighteen months. What changed is that my parent organisation now has a buyable, deployable, governable system to do what they have been asking their internal consulting and transformation teams to model. Until last week, when my group CFO asked whether we could collapse the finance and accounting tower from 1,400 people to 200 supervisors and an agent fleet, the answer was "in theory, on a roadmap, with the following risks". This week the answer is "Google sells the platform, Adobe and UKG have agents in the gallery, Anthropic's Claude is on the model garden, and the procurement team can issue the PO". That is not a small change. That is the whole change.

I have spent seven years building the capabilities of this captive. I have a strategy deck that says we are a 4,000-person centre of excellence with the parent. By the end of this calendar year, I have to write a new one that explains why we are still 4,000 people. I do not yet know what that deck says.

---

Q2: Your parent's investor relations team is on record that the captive is a strategic asset, not a back office. Has anything you have heard this quarter contradicted that publicly?

Publicly, no. Privately, the contradiction is everywhere. Three weeks ago, our group CFO asked the head of operations to model a 60 per cent reduction in our offshore finance and accounting tower over twenty-four months, with agentic AI carrying the load. He asked it as a stress test, not as a decision. But the stress test exists. The stress test did not exist eighteen months ago. The stress test will not stop existing.

There is a difference between the public position and the private modelling, and I have come to believe that the private modelling is what determines next year's budget. The investor relations deck describes our captive as innovation and value creation. The internal spreadsheet describes our captive as a line item that the agent architecture can compress. Both can be true at once. My job is to make the first one true faster than the second one closes the gap.

---

Q3: What are you actually doing right now, this month, that you were not doing six months ago?

Three things. First, I have stopped recruiting against the 4.25 lakh GCC hiring projection for FY26. The number is real, but the composition is wrong. I am no longer adding Commodities and Procedures roles regardless of business case. Every backfill in those two tiers now has to clear a two-quarter agent feasibility review before I approve the requisition. We were hiring 80 to 120 of these roles a month at the start of the year; we are now hiring 25 to 35.

Second, I have started a Grey Hair recruitment programme that I should have started in 2024. I am hiring senior people from the parent's consulting roster, from McKinsey and BCG India, and from Anthropic and Google partners' local teams, into a small unit that will sit between the agent platform and the business. They are expensive. I am not telling my CFO their salaries on a per-head basis.

Third, I have stopped pretending that internal upskilling will close the gap fast enough. We have rolled out our internal learning platform and the AI literacy modules; sixty per cent of the workforce has completed the foundation course. None of that produces an agent supervisor at the speed the parent is now moving. The honest answer is that some of my workforce will not make this transition, and I am not going to manage that by pretending otherwise.

---

Q4: What keeps you up at night?

Two things. The first is that I will be the executive who, when asked in 2028, has to explain why I did not see this coming when the architecture was visible in May 2026. I do see it coming. I am moving as fast as the parent's procurement, IT, and risk functions will let me. The risk function alone takes nine months to clear a new vendor for an agent that touches financial data. Nine months in 2026 is the entire useful life of an opportunity.

The second is the human one. Five thousand two hundred people work for this captive. They have mortgages in Hyderabad and Bengaluru, children in school, parents to support, families that moved cities to take these jobs. I am not going to lie to them that the architecture is not coming for a meaningful share of their roles. I am going to try to do right by as many of them as I can, which means moving the centre up the value chain faster than the architecture closes in. But I will not get all of them across. I knew that on Tuesday morning at 9am when the first Google Next 2026 announcement landed. I will know it for the rest of my career.

---

Q5: What do you wish the rest of the world understood about your position right now?

I wish the rest of the world understood that the executives running India's GCCs are not in denial. We see the architecture. We have read the Zinnov classification. We have run the Gartner numbers internally. We know that fifty-five per cent of our work sits in the tiers the platform is built to displace. We are not, as a community, telling our parents that the captives are safe; we are telling our parents that we can move up the stack with the right investment in the right window. That window is open right now and it will not be open in eighteen months.

The conversation that the global business press is having about us is the wrong conversation. The right conversation is not whether India's GCC story is over. The right conversation is whether the parent organisations will continue to invest in moving the captives up the stack, or quietly liquidate them and route the agentic work directly through Google Cloud and AWS. That decision will be made in board meetings in London and New York between now and Q4. The captives do not get a vote in those meetings. We have one chance to influence them, which is to show, between now and then, that we are the fastest path to the new operating model. If we do that, we keep the work and we keep most of the workforce, restructured. If we do not, the work moves to the platform and the workforce moves to LinkedIn.

---

Priya is one of six executives in similar roles I have spoken to in the past fortnight. Five of the six are running the same playbook she describes: pause Commodities-tier hiring, accelerate Grey Hair recruitment, and lobby the parent for an extended runway. None of them are confident the runway is long enough. The sixth has already begun a phased reduction. The next eight quarters will decide which playbook the rest of the industry follows, and the architecture that arrived this week has already foreclosed the option of not choosing one.