Summarize this blog post with:
India's GCC story has moved from scale to strategic consequence. With over 2117 GCCs employing more than 2.36 million professionals, the market is expected to reach around $110 billion by 2030.
But scale is not maturity.
That was the question Shiraz Mishra, Head - International Markets, Polestar Analytics, brought into the room as host of the inaugural BODHI GCC Roundtable.
A few GCCs have evolved to become the focal point of enterprise AI, data and decision-making for global organisations. But that evolution is uneven, and the gap between the most advanced GCCs and the rest is widening.
That widening gap set the stage for a candid conversation with Anirban Nandi, who leads data and AI globally at Albertsons Companies. Drawing from his experience across retail, enterprise AI, and GCC environments, Anirban offered a grounded view of what is really changing inside capability centres.
Anirban: Cost arbitrage! That may've been true 20 years ago, but it is increasingly disconnected from reality. Walk into any mature GCC today, or even a newer one, and you'll find teams owning AI platforms, data ecosystems, digital products, cybersecurity programmes, and business-critical decision-making processes. The talks have shifted from "how cheaply can this be done?" to "how much value can this team create?"
There's a deeper misunderstanding too. People assume GCCs are primarily an operating model or a location strategy. The best new-age GCCs are neither. They are talent and leadership strategies that happen to deliver cost efficiency along the way, not cost strategies that happen to deliver talent.
Anirban: The friction is real. But the challenge is not really about understanding the value of AI. The complexity is whether the organization can absorb the change needed to utilize it effectively.
In retail, decisions ripple through associates, customers, inventory, supply chain, pricing, compliance, and many other parts of the business. So, the question is never just "can we deploy AI?" It is "can we deploy AI responsibly and at scale while maintaining customer trust?"
The hardest part of AI transformation is not the model itself. It is redesigning the workflows, incentives, and decision-making processes around the model. That is where many AI efforts struggle. Organisations that succeed, focus less on isolated AI experiments and more on operational adoption.
Anirban: Five years ago, we hired people who could write great Python code. Presently, we're looking for people who can answer a very different question: should we even be writing this code in the first place?
Today, tech skills remain mandatory, but they're no longer sufficient. The people we value most are connectors. They can talk with business leaders, grasp the issues, translate it into a tech solution, and then drive adoption.
As AI automates parts of testing, coding, solution design, and documentation the constraints move. It is no longer just execution. It is defining the right problem, making trade-offs, and influencing stakeholders.
Anirban: The ultimate sign of ownership is when business leaders stop asking "can you build this?" and start asking "what do you think we should do?" I think of the journey as three stages of trust:
- Execution — Can they deliver what we asked for?
- Expertise trust — Do they understand the domain deeply?
- Judgement trust — Do we trust their recommendations?
Leaders have to create room for that ownership. It can't be demanded, it must be earned. That means giving teams responsibility of metrics, outcomes, and decisions, not just deliverables.
Teams also need to move from presenting solutions to presenting recommendations. Instead of saying, "Here is what we have built," they need to say, "Here is the problem we observed, here are the options we considered, and here is the approach we recommend."
Anirban: This question is top of mind in every boardroom right now. Historically, productivity was easy to measure. We counted reports generated, tickets closed, lines of code written, or projects delivered.
AI complicates all of that. If an engineer completes a task in 30 mins instead of 3 days because of AI assistance, have they become less productive as they produced fewer hours of work? Obviously not.
The focus has to move from activity metrics to outcome metrics: decision quality, decision speed, automation impact, business value, customer experience improvement, and risk reduction.
The trap, organisations fall into, is measuring AI with pre-AI metrics. The measurement framework itself has to evolve, and it has to be unique to each organisation's processes. There is no universal silver bullet. The framework depends on how the organization works, where AI is expected to create measurable value, and what it is trying to improve.
Anirban: Tokens are easy to count, and that is exactly why people reach for them. But impact and activity are not the same thing.
If one executive utilizes AI to generate 100,000 lines of code, and another executive asks a question that eradicates the requirement for that work altogether, who was more productive? I would say the second person.
The risk is that we become very efficient at producing more content, more code, and more analysis without creating proportionally more value. AI productivity has to be measured by impact, not usage volume.
Anirban: The most advanced GCCs will look very different. They will operate as globally integrated capability hubs rather than geographically distributed delivery centres.
The gap between leaders and laggards will not be measured by technology alone. It will be measured by how effectively GCCs integrate with the culture of the headquarters.
Culture is extremely important, especially for a new GCC. The ability to become part of the parent organisation's decision rhythm, priorities, and leadership expectations will separate GCCs that mature into enterprise capability hubs from those that remain delivery extensions.
The next phase of GCC maturity will be less about where the work sits and more about how decisions are taken.
The centres that come out as winners will be the ones that grasp the business deeply enough to challenge the brief, utilize AI responsibly enough to earn trust, and measure success through outcomes rather than activity.
That is where the real shift begins. A GCC is no longer valuable because it can execute more work. It becomes valuable when it helps the enterprise make better calls, move faster with confidence, and turn capability into measurable benefits.
The mandate is changing. Delivery will still matter, but judgement will matter more.