There is a diagram that exists, in some form, in every capability center. Sometimes it lives on a wall, sometimes in a process tool, sometimes only in the head of the one person who understands the whole flow. It maps an outcome the enterprise cares about, an order fulfilled, a claim settled, a product released, a month closed, across every team that touches it. Read the diagram left to right and watch the geography. A box in the center. A box at headquarters. Two boxes in the center. A box with a vendor. A box nobody has updated since the person who owned it moved on.
Now ask the diagram one question: who owns the outcome? Follow the boxes looking for a single name, a person who would be congratulated if the whole thing worked brilliantly and summoned if it failed. In most centers, for most outcomes, the search comes back empty. Every box has an owner. The line through the boxes has none. The center owns eleven pieces of the thing and none of the thing, and this is the third surface where the strain shows: the center delivers the pieces, and the whole is owned somewhere else.
Like the other surfaces, this one began as good judgment. Work reaches a young center in pieces because pieces are how sensible enterprises move work. Nobody transfers an end-to-end outcome to an unproven team on another continent. You move the reconciliation first, then the reporting, then a slice of the analytics, whatever is separable, documentable, and safe to hand over. Each piece arrives with its own sponsor, its own service levels, its own reporting line back to the function that sent it. This is the right way to start.
It is a limiting way to stay, and the limit arrives silently. Years pass. The pieces multiply, five becoming fifty becoming an operating landscape. What never happens, because no single transfer ever requires it, is the assembly: the moment someone looks at fifty accumulated pieces and notices that eleven of them, joined together, would constitute a whole outcome the center could simply own. The pieces were moved one sponsor at a time. Assembling them would take a decision that crosses sponsors, and decisions that cross sponsors belong to nobody, which is the recurring signature of every surface in this series.
So the center grows into a particular shape: teams organized as mirrors of the departments that sent them work, sitting under one roof, often on one floor, each reporting into a different function an ocean away. The industry data captures the raw material of this precisely. Nasscom and Zinnov found that almost ninety percent of India's centers now operate as multi-functional, running technology, operations, and product engineering side by side. Multi-functional is genuinely an achievement; it means the enterprise trusts the center with breadth. But breadth of functions and ownership of outcomes are different facts, and a center can hold the first for a decade without ever being handed the second. Many functions under one roof, assembled into wholes somewhere else, is precisely the pattern.
The phrase people reach for is silos, and it is half right. What makes a center's version distinctive is that the silos share everything except the work. Same building, same cafeteria, same town halls, same site leader in the same annual photograph. The proximity creates an expectation of coherence that the design never delivers, and the gap between the expectation and the fact produces a very specific set of daily experiences.
An outcome that takes four teams and belongs to none of them.
Teams named after the departments of a headquarters an ocean away.
New work bolted on first and connected later, if ever.
The last of those deserves a moment, because it is where the pattern renews itself. In a center organized by pieces, every new transfer of work repeats the founding move. A function at headquarters decides to move something, negotiates directly with a team it already knows, and the work lands wherever the relationship was, with its connections to everything else left as an exercise for later. Cross-border additions are the sharpest case: a second country site, an acquired team, a new charter, each arriving through its own door, each wired back to its sponsor, none wired to each other. I sat through the integration seasons that follow such arrivals, and the striking thing was never conflict. It was the absence of any forum where the question of how the new piece joined the whole could even be raised, because there was no whole for it to join.
Naming follows the same logic, and naming matters more than it appears to. Walk a floor and read the team names off the desks and the distribution lists. In a center of pieces, the names are borrowed: the team is called what the sending department is called, an ocean away, because the team was born as that department's extension. The names encode the design. A team named after someone else's function will think of itself as someone else's function, will escalate to that function, will measure itself in that function's terms. I have watched renaming exercises treated as branding, and they always fail as branding, because the name was never the problem. The name was the honest label on an arrangement, and the arrangement is what has to change first.
The costs of piecework are easy to underestimate because each one, alone, looks like ordinary friction.
There is the coordination tax: the standing meetings, the handoff documents, the status calls whose only purpose is to reconnect things the design separated. In a center of a thousand people organized by pieces, a meaningful share of everyone's week is spent stitching, and the stitching is invisible in any budget because it is smeared across every calendar. There is the accountability fog: when the outcome slips, four teams can each demonstrate that their piece performed, and the review ends the way such reviews always end, with an action item about communication. And there is the innovation ceiling, the least visible cost and the largest. Improvement happens at seams. A team that owns a piece can optimize the piece; only an owner of the whole can redesign the flow, remove a handoff entirely, or notice that three pieces should be one. A center without wholes can get better at everything and transform nothing.
There is a cost to the people, too. Talented professionals know the difference between running something and processing something, even when the org chart declines to acknowledge it. A center of pieces offers its best people a career of adjacency: near the product, near the decision, near the customer, owner of none of them. Some make peace with that. The ones you most want to keep usually do not, and their departures land in the attrition number, filed under compensation, unexamined.
One more honest observation belongs here, because leaders inside a center of pieces often sense the problem and reach for the wrong first move. The instinct is to create coordination: a program office, an integration layer, a weekly council of the four teams. These help, briefly, the way a brace helps. But coordination is the cost of missing ownership, and adding more of it institutionalizes the miss. The center ends up with a permanent apparatus for connecting things that should never have been separate, staffed by capable people whose entire role is compensating for the design. Whenever I find a large program office at the heart of a center, I read it the way a doctor reads a compensating posture: as the visible adaptation to an injury somewhere else.
The assembly, where it happens, tends to begin small and specific rather than sweeping. One whole, chosen because the center already runs most of its pieces. One owner, named out loud, with the seams redrawn around the outcome instead of around the sending departments. One year of running it that way, producing the thing no deck can produce: an existence proof, inside the enterprise's own walls, that the center can hold a whole. The second whole is always easier to ask for than the first. What the first one requires is only that somebody stops accepting the pieces as the natural unit of the center's life.
The industry has, in effect, built a public measure of this exact surface. Zinnov's maturity framework, the standard classification for India's centers, runs in four stages: Outpost, Satellite, Portfolio Hub, Transformation Hub. Strip away the labels and the ladder is a ladder of ownership. An outpost executes tasks. A satellite runs processes. A portfolio hub owns portfolios, wholes, end to end. A transformation hub owns change itself. By the 2025 landscape data, more than half of India's centers had climbed into the portfolio and transformation tiers, and the FY2026 report finds that nearly three-quarters of all centers have the potential to operate there within five years.
The engineering data adds a sharper edge to this. The 5-Year Journey report found engineering and R&D centers growing 1.3 times faster than the overall ecosystem, which is the market voting on where complexity is moving. Product and platform work is the least divisible kind of work there is; it resists piecework by nature, because a platform assembled across four unconnected teams stops being a platform. The centers winning that work are, almost by definition, the centers that have solved this surface. The flow of high-value work and the formation of wholes are the same current, moving together, in whichever direction a center's design allows.
Read those two numbers together and the strategic picture of this surface becomes clear. The ownership transition is not exotic; half the ecosystem has made it. And it is not automatic; the rest hold the potential and have not converted it. What separates the two groups, in my experience, is rarely capability and almost never intent. It is whether anyone ever performed the assembly: sat down with the map of accumulated pieces, drew the lines around the wholes hiding inside it, and then had the standing to ask the enterprise for them by name. Which is why this surface, like authority, so often traces back to the one the next essay takes up. A center is handed wholes when it is seen as a capability. A center seen as a cost is handed pieces, forever, one sponsor at a time.
The test for a leadership team is the one the diagram at the top of this essay proposes. Choose the three outcomes your center touches most. Follow each across the boxes. If you can name a single owner for each whole, this surface holds, and it is worth knowing why so you can protect it. If the search comes back empty, the finding is already made. The center did not fail to form wholes. It was never given any, and nobody has yet asked.
Nasscom & Zinnov, India GCC Landscape Report: The 5-Year Journey, 2025 (multi-functional centers; share of centers at portfolio and transformation maturity). nasscom.in
Nasscom & Zinnov, GCC Value Orbit: From Delivery Engine to Enterprise Nerve Centre, India GCC Landscape Report FY2026, July 2026 (maturity framework: Outpost, Satellite, Portfolio Hub, Transformation Hub; potential of centers to reach hub maturity). zinnov.com