Replacing the Worker
At a startup hub in an Indian city, developers are refining AI chatbots that can speak and text with strikingly human fluency.
The company, LimeChat, has an ambitious aim: to make traditional customer-service roles pretty much redundant. It claims its generative AI agents can cut by 80% the workforce needed to manage 10,000 customer inquiries each month. “Once you hire a LimeChat agent, you never have to hire again,” said Nikhil Gupta, the company’s 28-year-old co-founder.
India long served as the world’s back office, thanks to its inexpensive labor and widespread English fluency, but that advantage is now being disrupted. AI systems are rapidly taking over jobs in technical support, customer care, and data handling, forcing workers and employers to adapt. This shift is fueling demand for AI startups that promise lower staffing costs and greater scalability, even as many consumers still prefer to speak with a person. Based on interviews with 30 industry executives, recruiters, employees, and officials, it was found that India’s $283 billion IT sector is racing forward rather than slowing down, betting that new opportunities from AI will outweigh the losses. The gamble is significant: India could either prove that AI-driven disruption boosts developing economies or serve as a warning of its perils. Time will tell, although how much time is the question.
Changing Form or Eliminating Work?
The global conversational AI market is expanding by about 24% annually and is expected to reach $41 billion by 2030, according to Grand View Research. India, where IT contributes 7.5% of GDP, is leaning in. Prime Minister Narendra Modi has argued that technology doesn’t eliminate work, it simply changes its form. Yet critics warn that India isn’t prepared. “There’s no game plan,” said Santosh Mehrotra, a former government official and professor at the University of Bath.Business process outsourcing (which employs 1.65 million Indians) has seen hiring stagnate as automation accelerates. Net job growth in the segment fell from 177,000 in 2021–22 to fewer than 17,000 in each of the past two years, according to staffing firm TeamLease Digital. Workers describe a growing sense of insecurity as AI tools increasingly handle tasks once done by humans. One Bengaluru software employee earning $10,000 a year said she was laid off last month after her company adopted AI to evaluate sales calls. Some policymakers suggest expanding social protections to ease the transition, but others don't believe this is a threat.
Analysts say automation, combined with trade pressures and U.S. policy shifts, could cut India’s call-center revenue in half within five years. Still, some see opportunity: “India can move from the world’s back office to its AI factory,” said Pramod Bhasin, founder of Genpact.
LimeChat exemplifies this new frontier. Gupta says his company’s bots now handle 70% of customer queries and will reach 90–95% soon. Clients pay roughly the cost of three human agents to automate the work of 15. LimeChat’s revenue jumped from $79,000 in 2022 to $1.5 million in 2024, aided by a partnership with Microsoft Azure. Competitors like Haptik, owned by Reliance, offer similar AI agents at $120 each, claiming to cut support costs by 30%. At Mamaearth, a LimeChat client, AI assistants handle product questions, pregnancy-related recommendations, and even upset customers. “We can’t endlessly scale our support team,” said Vipul Maheshwari, the company’s head of product. The shift is visible across industries. Advertising firm The Media Ant reduced its staff by 40%, replacing sales and call-center teams with AI — including a voice bot named Neha that fields client calls in fluent English.
Yet not all experiments go smoothly. Some chatbots falter when asked complex questions, and many consumers still crave human contact: a 2024 EY survey found 78% of Indians prefer platforms that offer live support.ven so, India’s next generation is racing to adapt. In Hyderabad’s training centers, once focused on Java and Microsoft Office, students are now enrolling in AI and prompt-engineering courses.
