India's back office built the world's tech hub. Artificial intelligence is now shrinking it.

India Built the World's Back Office. A.I. Is Starting to Shrink It.

Steven Lee Myers reported from New Delhi, Gurugram, Ghaziabad and Mumbai. Paul Mozur reported from Taiwan. In Gurugram, the sprawling tech suburb outside New Delhi, Krishna Khandelwal is using artificial intelligence to build an army of chatbots designed to eliminate the kind of jobs that once lifted India into the ranks of the world's fastest-growing economies. Since last summer, his start-up, Hunar.AI, has offered companies bespoke A.I. voice agents that steer job applicants through virtually every step of the hiring process, from résumé screening to orientation. "For onboarding," he said in an interview in the company's headquarters in a shared work space, "you don't need humans at all." For a quarter century, India has made itself the world's back office, providing an educated, English-speaking work force to do tasks more cheaply than in the United States or Europe. The industry today employs more than six million people and is worth nearly $300 billion, more than 7 percent of the country's gross domestic product. Now, A.I. threatens to do to India what its outsourcing model did to the rest of the world: replace hundreds of thousands of office workers. Economies everywhere are bracing for an era in which A.I. tools automate entire categories of white-collar work, but the brunt could fall hardest on India, undermining two decades of effort to climb the value chain and establish a place in the global tech world. "It's a matter of time," said Deedy Das, a partner at Menlo Ventures, an investment firm that closely tracks A.I. "Markets are pretty efficient. If a tool exists that does a job cheaper, it will be adopted," he said. "I'm surprised it hasn't happened at a faster clip, but it will." The tremors are already being felt. Tata Consultancy Services, one of India's largest employers, has shrunk its work force to 580,000, a decline of more than 20,000 from a peak in 2022, when it hired 100,000 new workers in one year alone. Its main rival, Infosys, has also slowed hiring, while dozens of smaller start-ups laid off workers across the country in 2025, according to Inc42, a digital economy news outlet in India. Graduates of the country's universities and technical colleges are finding fewer openings, forcing them to scramble to "upskill," an increasingly popular term in the context of learning the A.I. technology that is reshaping the industry. Tech stocks in India were already slumping this year, but a speculative report on Feb. 22 by Citrini Research, an analytics company based in the United States, sent them spiraling, by painting a doomsday scenario about A.I.'s impact on India in particular. "The entire model was built on one value proposition: Indian developers cost a fraction of their American counterparts," the report said, imagining the world in the not-so-distant future of 2028. "But the marginal cost of an A.I. coding agent had collapsed to, essentially, the cost of electricity." Narendra Modi, India's prime minister since 2014, has recognized the challenge. Like many leaders, he has pledged to make the country an A.I. power, including by overseeing international deals and urging the country's own software engineers to develop new technologies and export them to the world. "There have been certain turning points that have shaped entire countries," Mr. Modi said during an international conference on A.I. in New Delhi last month, according to a translation of his remarks. "These turning points set the direction of civilization and transform the pace of development. Artificial intelligence is one such transformation in history." It is far from clear, however, whether India is positioned to make that transformation. While it has a highly educated work force, it lacks the infrastructure and natural resources needed to power A.I. products. At the A.I. conference in New Delhi, India's largest outsourcing firms announced deals with American companies including Anthropic and OpenAI to expand use of their products and to develop data center capacity to power them. The announcements, while welcomed as foreign investment in the economy, underscored how reliant the country remains on the United States for everything from the microchips to the foundational models that are driving the A.I. boom. Upheaval in the economy could undermine political support for Mr. Modi's government, whose party lost seats in the last parliamentary elections in 2024. The country already faces labor unrest that further job losses could worsen. Unemployment among educated young people reached as high as 65 percent of the total unemployed in 2022, according to a report by the International Labor Organization. While six million tech workers might not seem like a lot in a population of nearly 1.5 billion, they represent a politically vocal middle class concentrated in some of the country's most dynamic cities, like Bengaluru, Hyderabad and Pune. "If you're a young engineer getting out of university, I'd be worried," said Rajesh Nambiar, president of Nasscom, the tech industry association founded in 1988, when India's tech boom was in its infancy. "It's not going to be pretty out there." Mr. Nambiar and other industry leaders and experts are not exclusively gloomy. They point out that A.I. is already driving local start-ups in India that aspire to build products for global markets. "This is the best place to be," said Nikhil Gupta, a founder of LimeChat, an e-commerce company in Bengaluru. LimeChat designs chatbots for retailers on WhatsApp, the Meta-owned messaging app, which has more users in India than anywhere in the world. Mr. Gupta grew up in Cupertino, Calif., in the heart of Silicon Valley, but returned to India for high school and college and decided to build his start-up there. India, he said, "has all the ingredients already to provide services for the world." For India's outsourcing giants, deep relationships with multinationals and the ability to manage sprawling software systems mean that they will not disappear overnight. They have already expanded beyond call centers, offering services like human resources and accounting. In recent years, hundreds of international companies, many of them American, have opened their own operations in India, known as global capability centers, rather than simply contracting out the work. K. Krithivasan, the chief executive of Tata Consultancy Services, said the company planned to remake itself as an A.I. powerhouse. On an earnings call in January, he said revenues for the company's A.I. services had swelled in the last quarter to $1.8 billion, on an annualized basis. "We remain steadfast in our ambition to become the world's largest A.I.-led technology services company," he told investors. On the ground, however, the prospects seem grim. In call centers and back-office operations, hiring has slowed to a crawl. A.I. is creating new kinds of jobs -- including training A.I. chatbots by labeling data and teaching robots tasks like folding clothes -- but the jobs are repetitive, poorly paid and offer few paths upward. Young graduates can now make more money making deliveries for the country's booming e-commerce companies than they once could in entry-level tech, analysts and executives said. But for many, cultural expectations around careers have not caught up with the lowered prospects, said Siddharth Srivastava, a product leader at Instawork, a staffing platform based in San Francisco. "They don't have skills, but they have ambitions," he said. "They can't see their socioeconomic strata doing a blue-collar job." At the R.D. Engineering College in Ghaziabad, a gritty industrial city with an estimated 2.5 million people east of New Delhi, tech companies like Tata and Infosys once came calling in search of recruits. Now the college has to seek the companies out, according to Mohd Vakil, the dean of academics. Only a few years ago, 85 percent of graduates, on average, had jobs when they finished, said Mr. Vakil. In the last two years, it was 75 percent. With fewer companies hiring, competition has become fierce. During the semester break in January, Complygate, a background-check company based in Birmingham, England, held a 10-day training course on campus for nearly 100 students. The course ended with a test. Only the top scorers -- roughly 10 -- received job offers. "A.I. has changed the business model," said one student, Tathagat, 23, who goes by one name. A native of Bihar, a state bordering Nepal, Tathagat said the onus was on students to learn the skills that this new model demanded. "It depends on us," he said. "If you adapt with the times, it's a great time to graduate." The college has had to scramble to keep its curriculum up-to-date with technology, which changes as quickly as semesters pass. "You have to move with the markets," said Rakesh Sharma, the college's president. Referring to the students, he added, "We are making all efforts to make them employable." Hunar.AI offers a window into both the potential and the disruption. Mr. Khandelwal, who is 38, began by creating a traditional job-recruiting company with 65 employees handling calls to prospective applicants, mostly in banking but also for local Starbucks franchises. The company then fed more than four million minutes of recorded calls into a large language model using OpenAI's products to train its chatbots to make and conduct the calls in a robotic but fairly convincing conversational tone. Since launching last August, Hunar.AI has signed 70 clients and brought in $3 million in revenue. Its success illustrates the challenge facing India. Instead of 65 recruiters, the company now has 45. Mr. Khandelwal expects that number to dwindle to 25 or fewer, even as business grows. He estimated that the company has already eliminated 1,000 jobs in human resources. By the end of the year, he said, "we'll be doing the work of 10,000."

India built the world's back office. AI is starting to shrink it

India's IT outsourcing sector faces a significant challenge as artificial intelligence automates jobs. Companies like Hunar.AI are using AI to streamline hiring and onboarding. This shift threatens the employment of millions, impacting graduates and the nation's economic model. While the government aims to make India an AI power, reliance on foreign technology remains a concern. In Gurugram, the sprawling tech suburb outside New Delhi, Krishna Khandelwal is using artificial intelligence to build an army of chatbots designed to eliminate the kind of jobs that once lifted India into the ranks of the world's fastest-growing economies. Since last summer, his startup, Hunar.AI, has offered companies bespoke AI voice agents that steer job applicants through virtually every step of the hiring process, from resume screening to orientation. "For onboarding," he said in an interview in the company's headquarters in a shared workspace, "you don't need humans at all." For a quarter century, India has made itself the world's back office, providing an educated, English-speaking workforce to do tasks more cheaply than in the United States or Europe. The industry today employs more than 6 million people and is worth nearly $300 billion, more than 7% of the country's gross domestic product. Now, AI threatens to do to India what its outsourcing model did to the rest of the world: replace hundreds of thousands of office workers. Economies everywhere are bracing for an era in which AI tools automate entire categories of white-collar work, but the brunt could fall hardest on India, undermining two decades of effort to climb the value chain and establish a place in the global tech world. "It's a matter of time," said Deedy Das, a partner at Menlo Ventures, an investment firm that closely tracks AI. "Markets are pretty efficient. If a tool exists that does a job cheaper, it will be adopted," he said. "I'm surprised it hasn't happened at a faster clip, but it will." The tremors are already being felt. Tata Consultancy Services, one of India's largest employers, has shrunk its workforce to 580,000, a decline of more than 20,000 from a peak in 2022, when it hired 100,000 new workers in one year alone. Its main rival, Infosys, has also slowed hiring, while dozens of smaller startups laid off workers across the country in 2025, according to Inc42, a digital economy news outlet in India. Graduates of the country's universities and technical colleges are finding fewer openings, forcing them to scramble to "upskill," an increasingly popular term in the context of learning the AI technology that is reshaping the industry. Tech stocks in India were already slumping this year, but a speculative report on Feb. 22 by Citrini Research, an analytics company based in the United States, sent them spiraling, by painting a doomsday scenario about AI's impact on India in particular. "The entire model was built on one value proposition: Indian developers cost a fraction of their American counterparts," the report said, imagining the world in the not-so-distant future of 2028. "But the marginal cost of an A.I. coding agent had collapsed to, essentially, the cost of electricity." Narendra Modi, India's prime minister since 2014, has recognized the challenge. Like many leaders, he has pledged to make the country an AI power, including by overseeing international deals and urging the country's own software engineers to develop new technologies and export them to the world. "There have been certain turning points that have shaped entire countries," Modi said during an international conference on AI in New Delhi last month, according to a translation of his remarks. "These turning points set the direction of civilization and transform the pace of development. Artificial intelligence is one such transformation in history." It is far from clear, however, whether India is positioned to make that transformation. While it has a highly educated workforce, it lacks the infrastructure and natural resources needed to power AI products. At the AI conference in New Delhi, India's largest outsourcing firms announced deals with U.S. companies including Anthropic and OpenAI to expand use of their products and to develop data center capacity to power them. The announcements, while welcomed as foreign investment in the economy, underscored how reliant the country remains on the United States for everything from the microchips to the foundational models that are driving the AI boom. Upheaval in the economy could undermine political support for Modi's government, whose party lost seats in the last parliamentary elections in 2024. The country already faces labor unrest that further job losses could worsen. Unemployment among educated young people reached as high as 65% of the total unemployed in 2022, according to a report by the International Labor Organization. While 6 million tech workers might not seem like a lot in a population of nearly 1.5 billion, they represent a politically vocal middle class concentrated in some of the country's most dynamic cities, including Bengaluru, Hyderabad and Pune. "If you're a young engineer getting out of university, I'd be worried," said Rajesh Nambiar, president of Nasscom, the tech industry association founded in 1988, when India's tech boom was in its infancy. "It's not going to be pretty out there." Nambiar and other industry leaders and experts are not exclusively gloomy. They point out that AI is already driving local startups in India that aspire to build products for global markets. "This is the best place to be," said Nikhil Gupta, a founder of LimeChat, an e-commerce company in Bengaluru. LimeChat designs chatbots for retailers on WhatsApp, the Meta-owned messaging app, which has more users in India than anywhere in the world. Gupta grew up in Cupertino, California, in the heart of Silicon Valley, but returned to India for high school and college and decided to build his startup there. India, he said, "has all the ingredients already to provide services for the world." For India's outsourcing giants, deep relationships with multinationals and the ability to manage sprawling software systems mean that they will not disappear overnight. They have already expanded beyond call centers, offering services such as human resources and accounting. In recent years, hundreds of international companies, many of them American, have opened their own operations in India, known as global capability centers, rather than simply contracting out the work. K. Krithivasan, the CEO of Tata Consultancy Services, said the company planned to remake itself as an AI powerhouse. On an earnings call in January, he said revenues for the company's AI services had swelled in the past quarter to $1.8 billion, on an annualized basis. "We remain steadfast in our ambition to become the world's largest AI-led technology services company," he told investors. On the ground, however, the prospects seem grim. In call centers and back-office operations, hiring has slowed to a crawl. AI is creating new kinds of jobs -- including training AI chatbots by labeling data and teaching robots tasks like folding clothes -- but the jobs are repetitive, poorly paid and offer few paths upward. Young graduates can now make more money making deliveries for the country's booming e-commerce companies than they once could in entry-level tech, analysts and executives said. But for many, cultural expectations around careers have not caught up with the lowered prospects, said Siddharth Srivastava, a product leader at Instawork, a staffing platform based in San Francisco. "They don't have skills, but they have ambitions," he said. "They can't see their socioeconomic strata doing a blue-collar job." At the R.D. Engineering College in Ghaziabad, a gritty industrial city with an estimated 2.5 million people east of New Delhi, tech companies such as Tata and Infosys once came calling in search of recruits. Now the college has to seek the companies out, according to Mohd Vakil, the dean of academics. Only a few years ago, 85% of graduates, on average, had jobs when they finished, Vakil said. In the past two years, it was 75%. With fewer companies hiring, competition has become fierce. During the semester break in January, Complygate, a background-check company based in Birmingham, England, held a 10-day training course on campus for nearly 100 students. The course ended with a test. Only the top scorers -- roughly 10 -- received job offers. "AI has changed the business model," said one student, Tathagat, 23, who goes by one name. A native of Bihar, a state bordering Nepal, Tathagat said the onus was on students to learn the skills that this new model demanded. "It depends on us," he said. "If you adapt with the times, it's a great time to graduate." The college has had to scramble to keep its curriculum up-to-date with technology, which changes as quickly as semesters pass. "You have to move with the markets," said Rakesh Sharma, the college's president. Referring to the students, he added, "We are making all efforts to make them employable." Hunar.AI offers a window into both the potential and the disruption. Khandelwal, who is 38, began by creating a traditional job-recruiting company with 65 employees handling calls to prospective applicants, mostly in banking but also for local Starbucks franchises. The company then fed more than 4 million minutes of recorded calls into a large language model using OpenAI's products to train its chatbots to make and conduct the calls in a robotic but fairly convincing conversational tone. Since launching last August, Hunar.AI has signed 70 clients and brought in $3 million in revenue. Its success illustrates the challenge facing India. Instead of 65 recruiters, the company now has 45. Khandelwal expects that number to dwindle to 25 or fewer, even as business grows. He estimated that the company has already eliminated 1,000 jobs in human resources. By the end of the year, he said, "we'll be doing the work of 10,000." This article originally appeared in The New York Times.