字号·· | 护眼
美国消费者新闻与商业频道

从机房到大脑From the engine room to the brain

点「原文对照」整页切到原文,或双击某段只看那段的原文。

雷玛·滕杜尔卡的开场白从加巴尔·辛格说起。十年前,这位CNBC-TV18主播在索菲特BKC酒店大沙龙厅对在场人士说,每次关于印度全球能力中心(GCC)的对话,最终都会落到《复仇的火焰》中那个反派问手下的问题:“有多少人?”而答案——人数——多年来一直是一个中心重要性的衡量标准。她说,这场对话已经从成本套利迅速转向规模,现在又开始转向AI领导力以及知识产权和平台的所有权。

Reema Tendulkar opened with Gabbar Singh. A decade ago, the CNBC-TV18 anchor told the room in the Grand Salon at Sofitel BKC, every conversation about India’s global capability centres eventually arrived at the question the Sholay villain demands of his men, kitne aadmi the, and the answer, a headcount, served for years as the measure of how much a centre mattered. That conversation, she said, had moved quickly from cost arbitrage to capacity, and was now turning towards AI leadership and the ownership of intellectual property and platforms.

接着,她用一句话点出了当晚的愿景。“韩国有SK海力士,台湾有台积电,”她说,“但我们有GCC。”这场活动是科塔克马恒达银行呈献的印度GCC领导力对话孟买专场,由CNBC-TV18联合举办,于9月23日举行,是仅限受邀者参加的晚间活动,也是该系列六个城市专场之一。其邀请函直白地提出了问题:GCC的使命正在改变,印度准备好迎接接下来的一切了吗?

Then she gave the evening its ambition in a single line. “Korea ke paas SK Hynix hai, Taiwan ke paas TSMC hai,” she said, “par hamare paas GCC hai.”The occasion was the Mumbai edition of Kotak Mahindra Bank presents India GCC Leadership Dialogue, held in association with CNBC-TV18 on 23 September, an invitation-only evening and one of six city editions in the series. Its invitation posed the question plainly: the GCC mandate is changing, so is India ready for what comes next?

一位投资者解读GCC的故事。科塔克马恒达资产管理公司董事总经理尼莱什·沙阿以开幕致辞设定了背景,他首先把一个常见的抱怨变成了一个数据点。他对听众说,想象一下迪拜,脑海中浮现的是帆船酒店和安静整洁的街道;想象一下印度,浮现的是拥堵的交通——他离会场步行只需十分钟的办公室,却花了他半小时才到。

An investor reads the GCC storyNilesh Shah, Managing Director of Kotak Mahindra Asset Management, set the context with the opening address, and began by turning a familiar complaint into a data point. Picture Dubai, he told the audience, and the mind supplies the Burj Al Arab and quiet, clean streets; picture India, and it supplies traffic, and his own office, a ten-minute walk from the venue, had taken him half an hour to reach.

然而,他援引国际货币基金组织和世界银行的数据说,过去三十年里,迪拜经济大约增长了8.5倍,而印度增长了大约10.5倍;这种增长分布在14.5亿人中间,虽然更难察觉,但同样真实。2014年印度还是世界第十大经济体,如今按当天的卢比汇率换算已是第五或第六大,而在他看来,印度正迈向第三,GCC是最大的推动力之一。他引用的数字印证了这一判断。

Yet over the last three decades, he said, citing IMF and World Bank data, Dubai’s economy grew roughly eight and a half times while India’s grew roughly ten and a half; that growth spread across 145 crore people and is harder to see, though no less real. India was the world’s tenth-largest economy in 2014, is fifth or sixth today depending on the day’s rupee conversion, and, in his view, is headed for third, with GCCs among the biggest drivers. The numbers he cited made the case.

目前有2,100多家全球能力中心(GCC)雇用了近250万名专业人士,并在其周边创造了大约三到四倍的间接就业岗位。它们的营业收入约为500亿美元,但如果采用适当的转让定价基准,它们所创造的价值则接近98亿美元。他补充说,在过去十年里,全球能力中心在业务规模和价值方面均优于印度IT服务公司。他说:“对于共同基金投资者而言,不幸的是,你们当中没有一家是上市的。否则,我早就轻易卖掉我所有的IT公司,转而买入全球能力中心了。”

More than 2,100 GCCs now employ a little under 2.5 million professionals, with perhaps three to four times as many indirect jobs created around them. Their revenue rounds off to about $50 billion, but apply an appropriate transfer-pricing benchmark and the value they create comes closer to $98 billion. Over the past decade, he added, GCCs have outperformed Indian IT services companies in both volume and value. “Unfortunately for mutual fund investors, none of you are listed,” he said. “Otherwise, I would have easily sold all my IT companies and bought into GCCs.”AI, he argued, will become the core operating model of business as the internet once did, lifting productivity five to ten times.

他认为,人工智能将像当年的互联网一样,成为企业的核心运营模式,将生产力提升五到十倍。他讲述了一个企业家的故事:该企业家的机器依赖于全球仅有三家制造商生产的一个零部件,在售出哪怕一台机器之前,都需要获得其德国供应商的书面批准。借助人工智能,该供应商大约60%到70%的专有技术路线图已被破解,其余的研究则掌握在七家印度研究机构手中,这些机构目前正受到政府鼓励,将其研究成果与业界共享。

He told the story of an entrepreneur whose machines depend on a component made by only three manufacturers worldwide, its German supplier’s written approval needed before a single machine can be sold. With AI, some 60 to 70% of that supplier’s proprietary roadmap is already understood, and the remaining research sits across seven Indian institutes now encouraged by the government to share their work with industry.

这位企业家预计,在18个月内,他将尝试自己制造该零部件。沙阿将这一飞跃比作印度从固定电话向移动电话的跨越。他说,正如印度人当年学会了实施SAP的ERP系统和Oracle的数据库一样,他们现在能够构建全球将在大语言模型上运行的平台和应用程序,而全球能力中心将成为这项工作的核心。

Within 18 months, the entrepreneur expects to take a shot at making the component himself, a leap Shah likened to India’s jump from the landline to the mobile phone. Just as Indians once picked up the implementation of SAP’s ERP systems and Oracle’s databases, he said, they can now build the platforms and apps the world will run on large language models, with GCCs at the center of that work.

他说,全球能力中心发展历程的第一阶段是成本套利,第二阶段是能力建设,第三阶段则是人工智能领导力和价值创造。他说:“下一个十年将不再由全球能力中心的工作人数来定义。”

The first phase of the GCC journey, he said, was cost arbitrage, the second capacity building, and the third is AI leadership and value creation. “The next decade will not be defined by how many people work in GCCs,” he said.

“它将取决于从印度创造出多少创新、拥有多少源自印度的知识产权、设计出多少源自印度的全球产品,以及多少企业战略由印度塑造。”他告诉在座众人,通往“发达印度”的道路将穿过GCC(全球能力中心)的街巷。

一个席位,进而发声科塔克挑选的小组成员从同一栋大楼的不同楼层见证这段历程:Sudish Panicker,BNY基金与投资者解决方案全球运营主管兼印度区执行负责人;Rishi Dhuper,三菱UFJ养老金与市场服务印度中心董事总经理,该中心成立七年;Neha Shah,Investec全球服务印度公司首席执行官,该中心成立仅三年;以及Sachin Samant,科塔克马恒达银行银行与金融机构集团总裁兼业务负责人,他从银行家的角度近距离观察GCC发展。

“It will be defined by how much innovation is created from India, how much intellectual property is owned from India, how many global products are designed from India and how much enterprise strategy is shaped from India.”The journey to Viksit Bharat, he told the room, will pass through the lanes of the GCC.A seat, then a voiceKotak had chosen panellists who see that journey from different floors of the same building: Sudish Panicker, Global Head of Operations for Fund and Investor Solutions and Site Executive for India at BNY; Rishi Dhuper, Managing Director of the India Hub at MUFG Pension and Market Services, which was formed seven years ago; Neha Shah, Chief Executive Officer of Investec Global Services India, a centre barely three years old; and Sachin Samant, President and Business Head of the Banking and Financial Institutions Group at Kotak Mahindra Bank, who has watched GCCs up close from the banker’s side.

帕尼克更新了业界沿用多年的一个说法。过去的问题是GCC是否在决策桌旁有一个席位。“我认为对话现在已转向:GCC是否拥有同等的话语权?要让GCC留在桌旁而非仅仅露个面,需要什么条件?”他见过由印度主导整个AI战略的中心,也见过从损益表的成本端跨越到收入端、从印度业务发布全球产品的中心。部分优势源于结构:大多数组织纵向构建,而GCC是极少数让工程、产品、运营与风控横向并坐的场所之一。

Panicker updated a phrase the industry has used for years. The question once was whether GCCs had a seat at the table. “I think the conversation has now shifted to say, do GCCs have an equal voice? And what will it take for GCCs to stay at the table as opposed to just show up at the table?”He has come across centres where the entire AI strategy is led out of India, and others that have crossed from the cost side of the P&L to the revenue side, releasing global products from their India operations. Part of the advantage is structural: most organisations are built vertically, while a GCC is one of the few places where engineering, product, operations and risk sit together as a horizontal.

杜普尔通过与总部的对话描述了同一转变。五至七年前,对话围绕服务水平和人均混合成本。如今,它们转向集团的净资产收益率目标、AI采用情况,以及将一项5000万美元业务从一个平台迁移至另一个平台等战略项目。

Dhuper described the same shift through his conversations with headquarters. Five to seven years ago, they revolved around service levels and the blended cost of a headcount. Today they turn on the group’s return-on-equity targets, its adoption of AI, and strategic programs such as moving a $50 million business from one platform to another.

他解释说,成本套利之所以持续了大约十五年,是因为节省的成本每年都体现在损益表上,直到几乎没有什么可外派的了。到那时,决策者已经注意到,全球前2000大公司中约有800家已在印度设立机构,而他们在此招聘的人才带着“金标准”和最佳实践而来,准备度达70%至80%。“我们只需让他们了解我们的业务背景,他们就能上手运转。”萨曼特(Samant)这样表述:始于成本套利,现已演变为能力套利,涵盖产品开发、风险管理政策和企业变革。

Cost arbitrage, he explained, ran for a good fifteen years because the savings showed on the P&L every year, until there was little left to send. By then, decision-makers had noticed that some 800 of the world’s top 2,000 companies already had a presence in India, and that the talent they hired here arrived carrying gold standards and best practices, 70 to 80% ready. “We just have to give them context of our business, and they’ll be up and running.”Samant put it this way: what began as cost arbitrage has become capability arbitrage, spanning product development, risk management policy and enterprise change.

从第三阶段起步被问及行业在1到10分量表上的位置时,杜普尔(Dhuper)表示,老牌中心在前15至18年才实现的成果,新中心如今只需五年即可达成,这得益于一个从同伴互学、四大会计师事务所到NASSCOM(印度国家软件和服务公司协会)、DSCI(印度数据安全理事会)、银行和开发者的完整生态系统。他自己的中心成立七年,已处于成熟度的第三阶段。

Starting at phase threeAsked where the industry stands on a scale of one to ten, Dhuper said what older centres achieved in their first 15 to 18 years, newer ones now achieve in five, carried by an ecosystem that runs from peer-to-peer learning and the Big Four to NASSCOM, DSCI, banks and developers. His own hub, formed seven years ago, is already in the third phase of maturity.

他将三菱UFJ金融集团(MUFG)定位在“六七分偏上”,而其中心新增的每一项能力建设都始于一个问题:是否AI原生?他说:“AI不是事后诸葛亮。我们将从第一天起就将其融入服务交付模式。”内哈·沙阿(Neha Shah)的中心成立两三年,她表示,Investec受益于与前辈们的共同学习。在扩建过程中,AI原生性在每个决策中均居首位,角色因协同效应最强而迁移至印度,风险、运营和技术团队坐在一起办公。

He placed MUFG in “the higher sixes and sevens”, and every new capability his hub sets up begins with one question: is it AI native? “AI is not an afterthought,” he said. “We’re going to imbibe it in the service delivery model on day one.”Neha Shah, whose centre is two to three years old, said Investec had gained from learning alongside its predecessors. As it builds out, AI nativeness comes first in every decision, and roles move to India where the synergies are strongest, with risk, operations and technology sitting together.

AI已投入实战内哈·沙阿最清晰地描述了AI进入生产环节的情况。Investec于2024年初开始试验,2025年转入结构化提示工程,并向全球全体员工开放GitHub Copilot使用权限。过去向用户交付线框图的团队,如今构建可运行的原型,将两三个月的周期压缩至几周;一个用于跨地域成本分摊的财务智能体,已将三天的工作量缩减至数小时。

Where AI is already at workNeha Shah gave the clearest account of AI moving into production. Investec began experimenting in early 2024, moved to structured prompt engineering in 2025, and gave every staff member globally access to GitHub Copilot. Teams that once handed users wireframes now build live prototypes, compressing two to three months into weeks, and a finance agent for cost allocation across geographies has cut three days of work to hours.

她说,无论决策涉及控制、监管还是客户,AI都会提供信息支持,但最终由人做出决定。萨曼特称AI为增强工具,已在处理运营、交易处理以及分析支付中的反洗钱模式,而对欺诈调查或大额信贷决策的深度判断和最终决策仍由人类把控。他认为印度将在采用方面获胜,正如在数字支付领域所做的那样,特别是在银行和金融服务领域。

Wherever a decision touches controls, regulation or clients, she said, AI informs it, and a person makes it. Samant called AI an augmenter, already handling operations, trade processing and the analysis of payments for anti-money-laundering patterns, while deep judgment and final decisioning on fraud investigations or large credit decisions stays with people. He sees India winning on adoption, as it did in digital payments, especially in banking and financial services.

Kotak自身的部署便印证了这一点。Kompanion将Kotak的政策和流程说明放入内部助手,供全国一线员工关系经理在销售时使用。同样,AI智能体也被用于软件开发、信贷承销、客户服务及其他用例。

Kotak’s own deployments illustrate it. Kompanion places Kotak’s policies and process notes inside an internal assistant for frontline staff relationship managers selling across the country. Likewise, AI agents are being leveraged for software development, credit underwriting, customer service, and other use cases.”

杜普尔刚从对微软、亚马逊、CrowdStrike、OpenAI和Anthropic的考察之旅归来,表示他带回的问题比答案多,并总结出四大主题:信任、自上而下与自下而上相结合的采用模式、建立在“信任但要核实”银行家法则之上的治理,以及为每个人都会犯的错误提供遮护。“AI不关乎技术。AI关乎心态。关乎领导力,”他说。

Dhuper, back from a study tour of Microsoft, Amazon, CrowdStrike, OpenAI and Anthropic, said he returned with more questions than answers and four themes: trust, a mix of top-down and bottom-up adoption, governance built on the banker’s rule of trust but verify, and air cover for the mistakes everyone will make. “AI is not about technology. AI is about mindset. It’s about leadership,” he said.

衡量关键指标潘尼克指出,心态是行业的首要制约。他说,GCC(全球能力中心)建立之初是为了成为组织的引擎室。“我认为对我们而言,机遇与转型的挑战在于成为组织的大脑。”第二是衡量。行业长期以来将规模等同于人数,他称这一等式已失效,更好的指标应捕捉知识产权、创造的收入和业务成果。“那就是将技能、语境和成果结合在一起的地方。”

Measuring what mattersPanicker named mindset as the industry’s first constraint. GCCs were built to be the engine room of their organisations, he said. “I think the opportunity and the pivot challenge for us is to become the brain of the organisation.”The second is measurement. The industry has long equated scale with headcount, an equation he called broken, and better metrics would capture intellectual property, revenue generated, and business outcomes. “That’s where you bring skill, context, and outcome together.”

杜普尔给出了界定机遇的数字。印度GCC约250万人中,仅六千至七千人担任全球角色,约占劳动力的0.3%至0.4%,他说当这一比例达到3%时,行业将经历重大转型。

Dhuper supplied the number that frames the opportunity. Of roughly 2.5 million people in India’s GCCs, only six to seven thousand hold global roles, around 0.3 to 0.4% of the workforce, and when that share reaches 3%, he said, the industry will go through a significant pivot.

“我觉得前路广阔,未来光明,”他说。“这将会发生。这是不可避免的。”内哈·沙阿(Neha Shah)追溯了这一差距的根源,认为该行业最初作为交付中心,留下了供应商的心态,因此她将业务背景(context)设为首要任务,教导员工不仅要问做什么,还要问为什么。她说,总部也在发生转变,开始将印度视为一个既能设计工作又能交付工作的地方。

“I feel the runway is long, the future is bright,” he said. “It will happen. It’s inevitable.”Neha Shah traced the gap to the industry’s origins as delivery centres, which left a vendor mindset, and has made context her priority, teaching her people to ask why as well as what. Headquarters are shifting too, she said, coming to see India as a place that can design work as well as deliver it.

“不要在没有业务背景的情况下转移工作,”她说。“不要在没有决策权的情况下转移工作。”下一步最后,滕杜尔卡询问是什么能把全球前2,000家公司中剩下的1,200家吸引到印度。萨曼特呼吁进一步简化设立流程,并在中央和各州的鼓励基础之上,培养能够胜任更高级任务的人才。内哈·沙阿指出,需要进行能力建设,并培养人工智能时代任务所需的批判性思维和判断力,同时将学术界作为合作伙伴。

“Do not move work without context,” she said. “Do not move work without decision-making authority.”What comes nextTo close, Tendulkar asked what would bring the remaining 1,200 of the world’s top 2,000 companies to India. Samant called for making set-up ever easier, building on encouragement from the centre and the states, and for producing talent equipped for far more advanced mandates. Neha Shah pointed to capability building and the critical thinking and judgment AI-era mandates demand, with academia as a partner.

杜珀表示,商业领袖需要同时具备商业背景、商业敏锐度、人工智能与数字技能以及对监管的理解,并希望人工智能现在能够加速这一进程。

Dhuper said a business leader needs business context, commercial acumen, AI and digital skills and a grasp of regulation together, and hoped AI would now accelerate that journey.

综上所述,当晚勾勒出了清晰的前进道路。印度全球能力中心(GCC)故事的第一阶段建立在成本基础之上,第二阶段建立在产能基础之上;正如尼莱什·沙阿所言,第三阶段将建立在人工智能领导力和价值创造的基础之上。小组讨论列出了这一阶段的要求:在企业战略中拥有与席位相匹配的话语权,采用计算知识产权和成果而非员工人数的指标,从第一天起就在每项新能力中融入人工智能且仍由人类做出重要决策,以及由印度承担更大份额的全球角色。

Taken together, the evening sketched a clear road ahead. The first phase of India’s GCC story was built on cost and the second on capacity; the third, as Nilesh Shah put it, will be built on AI leadership and value creation. The panel set out what that phase demands: a voice in enterprise strategy to match the seat at the table, metrics that count intellectual property and outcomes instead of headcount, AI built into every new capability from the first day with people still making the decisions that matter, and a far larger share of global roles held from India.

随着全球2000家顶级公司中已有约800家落户于此,且按杜珀(Dhuper)的话说,发展空间依然广阔,未来十年印度全球能力中心(GCC)的发展将取决于它们所拥有的资产和发挥的主导作用,而不再那么依赖人数的多寡。活动在加甘迪普·阿胡贾(Gagandeep Ahuja)——消费者分销西区、东区及三州地区总裁兼地理主管,以及维沙尔·萨赫德瓦(Vishal Sachdeva)——跨国公司与新时代银行业务主管的发言中落下帷幕,两人阐述了科塔克马欣达银行(Kotak Mahindra Bank)如何在其整个生命周期中为全球能力中心提供支持:从第一天起就提供带有数字交易银行业务的协助开户服务、用于股权注入和外部商业借款的外汇管理法(FEMA)咨询服务台、用于套期保值的外汇业务台、通过吉夫特城(GIFT City)开设的外汇账户,以及分别针对应届毕业生、职场中期专业人士和高级领导层设计的员工银行服务。

With some 800 of the world’s top 2,000 companies already here, and the runway, in Dhuper’s words, long, the next decade of India’s GCCs will be measured by what they own and shape, and far less by kitne aadmi the. The evening closed with Gagandeep Ahuja, President and Geography Head, Consumer Distribution, West, East and Tri-State, and Vishal Sachdeva, Business Head, Multinationals and New Age Banking, setting out how Kotak Mahindra Bank supports a GCC through its entire life: assisted account opening with digital transaction banking from the first day, a FEMA advisory desk for equity infusion and external commercial borrowings, a forex desk for hedging, foreign currency accounts through GIFT City, and employee banking designed separately for graduates, mid-career professionals and senior leaders.

该行近30%的分支机构网络、即超过770家分行,分布在全球能力中心聚集的六大市场中,紧邻那些将构建未来发展的人士。

Nearly 30% of the bank’s branch network, more than 770 branches, sits in the six markets where GCCs cluster, close to the people who will build what comes next.

阅读媒体原文 ↗