Call centres, once the engine room of India's BPO exports, are evolving too.
Depending on the complexity, 30 to 50 per cent of voice and chat volumes are now handled by conversational AI.
The $54 billion business process outsourcing (BPO) industry -- long known for its human capital advantage and operational scale -- is undergoing a transformation more profound than any seen since its start almost three decades ago.
This time, the disruption is not being driven by headcount, but by algorithms.
Artificial intelligence (AI) and its more autonomous cousin, Agentic AI, which can act independently and make decisions without constant human intervention, are rapidly reshaping the architecture of global outsourcing, redefining not only how services are delivered, but what services are even needed.
It's not about replacing humans -- it's about augmenting human capabilities.
"We are moving beyond generic AI to building domain-specific intelligent AI solutions," said Sanjeev Vohra, chief technology and innovation officer at Genpact, the $4.77 billion business and technology services major.
"We are integrating domain-specific intelligent agents into business operations that adapt, learn and optimise in real time," he added.
From cost arbitrage to AI edge
BPOs have been synonymous with cost-cutting strategies of global enterprises.
Companies, spanning Fortune 2000 majors including GE, Amex, Bank of America, Citibank, JP Morgan, Prudential, Capital One, Pfizer, and hundreds of others, have banked on low-cost, high-quality talent from India to deliver a gamut of services -- from customer support to account reconciliation, from analytics to mortgage processing.
Now, this industry of 1,000-plus companies, which Nasscom estimates is worth $54 billion and employs around 2 million, is morphing into intelligent AI-driven cognitive ecosystems.
Automation, analytics, and empathy now coexist in tightly integrated systems.
Routine processes -- from invoice reconciliation to customer onboarding -- are being handled by intelligent bots that require little to no human supervision.
Birendra Sen, president of Business Process Services at Tech Mahindra, calls it a "tectonic shift". At $6.26 billion, Tech Mahindra is among the top six technology services providers.
"Up to 60 per cent of agent workloads, depending on the industry, are now managed by AI systems," Sen said.
"Call summarisation, sentiment analysis, and predictive insights are handled in milliseconds. This frees human agents to focus on what machines still can't do -- handle emotions, persuasion, and nuanced problem-solving."
From response to reasoning
Unlike traditional AI, which reacts based on rules or training data, agentic AI acts. It can assess a situation, make decisions, and take independent steps to resolve problems -- capabilities that are now being baked into BPO processes.
Take, for instance, Genpact's AP Capture, an agentic AI solution. Designed for accounts payable functions, the system uses GenAI agents to read, interpret, and reconcile invoices across large enterprise systems.
Initial deployments have cut manual processing efforts significantly, resulting in faster payments and lower error rates.
The firm also helped a major global investment bank tame a sprawling 600-million-node knowledge graph (a kind of database comprising multiple entities).
Using AI agents with reasoning and memory capabilities, the bank can now interpret complex user queries and navigate vast ontologies using natural language, turning dense datasets into strategic assets.
"We built a scalable AI agent system that leveraged perception to understand user queries," Vohra said.
"By making complex knowledge graphs accessible through natural language, we have empowered the bank to transform data complexity into a strategic advantage."
Call centres, once the engine room of India's BPO exports, are evolving too. Depending on the complexity, 30 to 50 per cent of voice and chat volumes are now handled by conversational AI, according to data from BPO firm Quatrro and analyst firm EY India.
"We've seen massive gains in automation for routine queries -- banking transactions, airline bookings, or mortgage processing," said Raman Roy, chairman and managing director, Quatrro BPO Solutions, a Gurugram-based business services provider.
"BPO in India is moving up the value chain. It is increasingly AI-assisted."
Agent-assist tools now 'listen' to live calls, pull up contextual data, and coach agents in real time -- cutting average handling times and increasing first-call resolution.
Yet, companies are proceeding with caution when it comes to voice AI.
"Depending on client maturity and process type, automation may account for 30-50 per cent of voice and chat volumes. But priority remains in maintaining human empathy and contextual judgement where it matters most," said Keshav R Murugesh, group CEO at WNS, a Mumbai headquartered BPO services provider.
While Tech Mahindra is seeing an AI-led transformation that is both technical and cultural, Sen describes their BPS unit as becoming "phygital" -- where physical human effort and digital intelligence combine.
For a financial services client, Tech Mahindra deployed emotion-aware AI and real-time speech analytics to scale, from reviewing customer interactions manually to 100 per cent automated.
"The return on investment is real," Sen said. "In low-complexity environments, we've seen up to 3x cost savings. But the bigger prize is in augmenting human intelligence, not replacing it."
Looking ahead, Sen envisions decentralised, on-demand contact centres powered by cloud and edge computing -- with agents and AI working as partners. "The future isn't man versus machine," he said. "It's man with machine."
For the $1.31 billion WNS, GenAI is no longer just a capability -- it's a revenue stream.
According to Murugesh, 5 per cent of the company's 2024-2025 (FY25) revenue came directly from GenAI-powered solutions.
The company recently helped a global insurance client identify claims recovery opportunities, reducing false positives by 60 per cent using AI models embedded with domain-specific rules.
In the healthcare space, AI systems have expedited diagnoses, reduced clinical errors and are generating automated medical summaries. In travel, WNS uses AI to personalise itineraries.
Internally, WNS has upskilled 22,000 employees in GenAI competencies. Across BPO providers, AI is not an add-on anymore.
It's baked into how they design customer experiences, automate workflows and drive business outcomes.
Industry veterans stress that this is not a transition to machines, but a convergence with them.
"AI excels at scale, speed, and precision," said Roy of Quatrro. "But complex decision-making, ethics, and context are still very human domains."
A collaborative man-machine future
Preeti Anand, Partner at EY India, agrees: "We're not seeing the end of human contact centres -- but an evolution."
EY estimates that up to 50 per cent of call types are ripe for automation via AI or chatbots, but another 20 to 30 per cent are best addressed by AI-assisted humans who can act on real-time data and insights.
"Human agents will continue to handle high-value, high-empathy tasks," Anand said, but they will do so with powerful AI co-pilots.
Despite fears of an AI-led demise of human jobs, most leaders see a more collaborative future. "AI will do the heavy lifting," said Roy. "But humans will still provide judgment and creativity."
In fact, a lot of new jobs are emerging. Nasscom's Senior Vice President Sangeeta Gupta said, "At present, the industry is expanding because of AI. And at least in the next 3 to 5 years, it will lean on more human talent.
"For instance, there is a lot of job creation happening in data annotation (labelling data) tasks."
Nasscom sees a 6 to 7 per cent annual growth rate for BPO in this fiscal.
According to the industry body, as AI becomes more embedded in enterprise workflows, entirely new job roles are emerging.
Business functions are seeing the rise of roles such as data scientist, AI strategy director, and Chatbot HR specialist.
In customer service, positions like AI conversation designer and virtual assistant trainer are becoming critical to managing AI-human interactions effectively.
Roy said, "Organisations that combine AI's capabilities with human insight will be the ones that thrive."
The BPO sector, which once rode a wave of global labour arbitrage, is now scripting its second act -- one where the next big outsourcing engine might not speak with an accent, but solve problems with agentic intelligence and clarity.
India, long the world's backoffice, is now positioning itself as the frontrunner in building the future of intelligent operations.
The co-pilots
- AI and humans work as partners
- Processes that involve a set of standard rules, like determining insurance premium, loan eligibility or mortgage processing, are done by both AI and humans
- Routine processes like invoice reconciliation, and customer onboarding are being handled by AI
- AI-led automation is key as teams upgrade to more complex tasks
Feature Presentation: Ashish Narsale/Rediff