What a global capability centre is today
A global capability centre is a unit an enterprise owns, or controls through a partner, that builds and runs capabilities for the wider organisation: engineering, cloud and reliability, data and AI, finance and audit, operations. Early centres were set up mainly to reduce cost. The centres that leadership teams now set up are expected to own outcomes: run platforms to agreed service levels, build products, keep controls audit-ready and bring scarce skills the headquarters cannot hire fast enough.
That shift changes what success depends on. Cost still matters, but it follows from the operating model, the leadership and the governance, not from the payroll alone.
Is a capability centre right for you?
A capability centre tends to pay off when most of the following are true:
- The work is long-running and core to the business, not a one-off project.
- You need specialist skills at a scale your current locations cannot supply, such as SRE, platform engineering, AI or qualified audit professionals.
- You want ownership of outcomes and of the knowledge built up, rather than buying hours from a vendor.
- Leadership is willing to give the centre real mandates, decision rights and visibility.
If the need is short-term, or the work cannot be clearly scoped and owned, a managed service or a project-based partner is usually the better fit.
Choosing the model
There are four common ways to stand up a centre. They differ mainly in how much you invest up front, how quickly the centre is productive and how much control you hold from day one.
| Model | How it works | Best for | Watch out for |
|---|---|---|---|
| Captive build | You set up your own entity, hire leadership and teams, and run everything yourself. | Large enterprises with in-country experience and a long horizon. | Longest time to productivity; you carry all setup risk. |
| Build-operate-transfer (BOT) | A partner builds and runs the centre under your governance, then transfers people, processes and assets to your entity on agreed terms. | Organisations that want ownership in the end but speed and lower risk at the start. | Transfer terms, timing and knowledge retention must be agreed at the outset. |
| GCC as a service | A partner provides a dedicated, branded capability with its own people, premises and compliance, run to your standards and KPIs. | Mid-size companies and first-time entrants that want a dedicated team without setting up an entity. | Make sure the team stays dedicated, and the contract keeps a path to transfer. |
| Hybrid | Start as a service or BOT for one function, then insource or expand as the centre matures. | Most organisations that are new to capability centres. | Needs a clear roadmap, or it drifts into a permanent vendor arrangement. |
For most first-time and mid-size entrants, the practical path is to start with one function under a GCC-as-a-service or BOT model, prove the operating model, then decide whether and when to own it outright. We compare these two models in more depth in GCC as a service or build-operate-transfer.
Choosing the location
India hosts the largest concentration of global capability centres, spread across established hubs such as Bengaluru, Hyderabad, Pune, Chennai, the Delhi NCR region and Mumbai, with newer centres opening in tier-2 cities. Whichever country and city you consider, compare them on the same factors:
- Depth of the specific skills you need, not general headcount. Senior SRE, cloud security and qualified audit professionals are scarcer than general developers everywhere.
- Time-zone overlap with the teams the centre will work with most closely.
- Infrastructure and business continuity: office options, connectivity and resilience.
- Government policy and incentives, which several Indian states now offer specifically for capability centres.
- Total cost: salaries, premises, attrition and management overhead together, not salaries alone.
Designing the operating model
The operating model decides whether the centre owns outcomes or simply executes tasks. Settle four things before you hire at scale:
- Mandate. Which services or processes the centre owns end to end, and which it supports.
- Decision rights. Who approves architecture, releases, changes to controls and hiring. Write it down as a responsibility matrix.
- Service levels and KPIs. Agreed before go-live, measured from the first month.
- Governance cadence. A weekly operational review, a monthly service review against KPIs and a quarterly executive review of value, maturity and the roadmap.
A five-stage setup roadmap
- Assess. Review the current state of the function you are moving or building: processes, controls, volumes, systems, skills and risks. Output: a current-state report, a risk register and a maturity baseline.
- Design. Define the target operating model: scope, ownership matrix, procedures and controls, team shape, tooling, KPIs and service levels. Output: a signed-off design and transition plan.
- Build. Hire leadership first, then the team. Run knowledge transfer, shadowing and reverse shadowing, document procedures and finish with a supervised parallel run. Output: go-live approval against agreed criteria.
- Operate. Deliver to the agreed service levels, with the governance cadence above and transparent reporting.
- Optimise. Work through a prioritised improvement backlog: automation, standardisation, stronger controls and a wider scope as the centre earns trust.
Each stage should end with an exit gate, so leadership signs off before the next one starts. This is the method we use on every engagement; see our approach.
Talent and leadership
Hire the centre's leaders before the teams. A strong site leader and practice leads who have run similar functions shape culture, hiring quality and credibility with headquarters. Senior-led teams also retain knowledge better: when expertise sits in documented procedures and shared ownership rather than in a few individuals, the centre survives attrition, which is a constant in competitive talent markets.
Compliance and risk
Plan for these from the design stage, with your legal, tax and security advisers:
- Data protection. Personal data moving from the EU or UK to another country needs a lawful transfer mechanism such as Standard Contractual Clauses. In India, the Digital Personal Data Protection Act, 2023 also applies.
- Information security. Most enterprise clients expect controls aligned to frameworks such as ISO 27001 or SOC 2, with access control, logging and incident response from the first day.
- Tax and transfer pricing. How the centre is paid for affects tax in both countries; take specialist advice before you finalise the model.
- Employment and continuity. Local labour law, notice periods and a backfill plan for critical roles.
Measuring success
Agree a small, balanced set of measures before go-live and review them monthly:
- Delivery: throughput, cycle time and on-time delivery against plan.
- Reliability: SLO attainment, incident recurrence and time to restore for the services the centre runs.
- Quality and control: exception rates, audit findings and rework.
- Cost: cost per unit of work or per service, not headcount cost alone.
- Capability: movement on a maturity scale for each function the centre owns.
Five common mistakes
- Treating the centre as a cost-cutting exercise, with no real mandate.
- Hiring teams before leaders.
- Moving work without documenting it, so knowledge never transfers.
- Agreeing KPIs after go-live instead of before it.
- Giving headquarters no visibility, so trust and scope never grow.
Frequently asked questions
How long does it take to set up a global capability centre?
It depends on the model and scope. A first function delivered through a GCC-as-a-service or BOT partner can be productive in months, while a full captive build, with its own entity and premises, takes considerably longer.
What is the difference between a GCC and outsourcing?
In outsourcing, a vendor delivers a service and owns how it is done. In a capability centre, the people and processes work as part of your organisation, under your governance, and the capability and knowledge stay with you.
Can a mid-size company set up a GCC?
Yes. Mid-size companies are among the fastest-growing groups of new entrants, usually starting with one function under a GCC-as-a-service or build-operate-transfer model rather than a full captive build.
Which functions should move first?
Start with a function that is well understood, measurable and important enough to matter, such as platform reliability, a product engineering stream or a finance process like accounts payable. Prove the operating model there before adding scope.
Who should own the centre inside the organisation?
A named executive sponsor at headquarters, with a site leader in the centre who reports into them. Shared ownership without a single accountable sponsor is a common reason centres stall.