We will need to learn how to validate, provide the right judgement, redirect, work iteratively, and test and learn.
May 8, 2026 | Hemant Kumar
Most leaders herald the promise of AI. Many employees, meanwhile, see it as a looming threat, a modern echo of the familiar taunt: “Anything you can do, I can do better.”
In India, this anxiety is especially charged. From IT services campuses in Bengaluru to back-office operations in Hyderabad and fintech hubs in Pune, entire industries built on the proposition that skilled Indian talent delivers quality work at scale are now asking, ‘What happens when an agent can do that work overnight, without a salary band or an annual appraisal?’
The question is not whether AI will change work. It already has. The real question is how leaders build constructive partnerships between humans and AI agents, and what that means for an economy of 1.4 billion people, with a workforce of nearly 600 million, entering the most consequential technological transition of our lifetimes.
McKinsey’s Global Institute research makes a striking opening claim: based on currently proven technologies, more than half of current work hours could be automated today. But reading this as a job-loss forecast misses the point entirely.
50%+ of current US work hours could be automated with today’s proven AI technologies, and the picture in India’s high-volume service sectors is no different.
“We shouldn’t lose sight of the fact that human beings are vital,” notes MGI partner Anu Madgavkar in the research. “Almost half of that work is beyond the capabilities of today’s technology. A lot of this work is cognitive, social, emotional, and interpersonal.”
Take the case of the BFSI industry in India. HDFC Bank, ICICI, and Kotak Bank handle thousands of loans, KYC documents, and queries every month. Verification of documents, scoring customers based on standard criteria, and initial queries are all within the ambit of AI technology.
Relationship banking, where a banker knows three generations of a family and a relationship officer structures a farm loan for a client during a difficult monsoon year, remains irreducibly human.
In a similar vein, code generation, test writing, and documentation, which take up huge chunks of developer time, are all being handled by AI coding assistants at TCS, Infosys, and Wipro. What happens to the millions of hours released from these tasks for the two million IT employees in India?
The most important insight from the research is that the vast majority of skills are shared, skills that AI agents and robots can bring to the workplace but that human workers also use. This is precisely the source of unease.
It means workers cannot stay at the same base level and expect to add value. It demands what Madgavkar calls “super-skilling” using AI to get better at the skills you already have.
“We can’t stay at that same base level on each skill and expect to add a lot of value. We need to enable everyone in the workforce to work with AI and get better at those skills.”
— Anu Madgavkar, MGI Partner, McKinsey & Company
The research identifies a core cluster of high-value human contributions likely to remain durable over the next five to ten years: critical thinking, negotiation, conflict management, team leadership, and complex problem-solving.
In many of these, humans will use AI as a lever, but not be substituted by it.
The radiology parallel offered in the McKinsey research is instructive: rather than displacing radiologists, AI diagnostic tools increased the effective supply of radiological expertise, unlocked latent demand, and grew the field.
Healthcare:
India has approximately 1 radiologist per 100,000 people, one of the lowest ratios in the world. AI-assisted diagnostic tools like those being piloted at AIIMS Delhi and integrated into platforms like Niramai (for breast cancer screening) do not replace the radiologist; they allow one radiologist to do the work of ten, reaching tier-2 and tier-3 cities that previously had no access at all.
Agriculture:
PM-Kisan and digital crop advisory platforms already push personalised guidance to over 110 million farmers. As agentic AI layers onto these platforms, the agricultural extension worker becomes a supervisor of AI-driven advisory agents, handling the exceptions, the conflict cases, and the distress calls that require a human voice.
Social-emotional skills, coordination, and process management will command a premium.
McKinsey’s Alexis Krivkovich puts it directly: as AI takes certain tasks off the human plate, organisations will want people spending more of their time on the skills that remain distinctly human.
And leaders will need something new in their personal repertoire: what the research calls “a voracious learning mindset.”
The research estimates that AI could generate almost 3 trillion dollars in annual value globally by 2030. Yet most organisations are not seeing meaningful gains at the enterprise level.
The pilots are live; the transformation is stalled.
The gap lies in how companies deploy. Most remain in what Krivkovich calls “pilot and point” mode, experimenting in isolated pockets, optimising one process at a time, generating intriguing dashboards but not EBITDA points.
The real opportunity lies in at-scale bets that cut across multiple leaders’ domains: supply chain connecting into customer service, connecting into manufacturing, connecting into finance; all reimagined together.
The Reliance Industries model is a useful reference point. Jio’s rollout was not a pilot. It was a total reimagination of the telecom and digital commerce stack simultaneously, underwriting mobile internet access for 400 million new users while building the retail and payments layer on top.
The lesson for Indian conglomerates and mid-sized enterprises today is the same: the companies that will capture AI’s value are those willing to place a few large, coordinated bets rather than scatter resources across 40 unconnected experiments.
For PSUs, the challenge is governance. An organisation like the Railways, processing over 23 million passengers per day with ticketing, logistics, freight, and maintenance running in parallel, has enormous untapped AI potential, but capturing it requires leadership alignment across silos that rarely speak to one another.
Madgavkar’s framing is the T-shaped capability model.
Horizontal capability building is necessary; everyone needs baseline fluency with AI tools. But the transformational value comes from a few deep vertical bets: reimagining end-to-end processes, not just adding AI tools to existing workflows.
For Indian businesses, this means asking not “Where can we use a chatbot?” but “Which part of our core profit pool is most vulnerable to disruption, and what does a fully reimagined version of that process look like?”
“Speed is a strategy in and of itself. The biggest risk companies face is waiting for more clarity before they make bets.”
— Alexis Krivkovich, Senior Partner, McKinsey & Company
As AI agents enter the workforce in earnest, the nature of management itself shifts.
The McKinsey research draws a sharp picture: agents work continuously, learn in real time, and sometimes outperform humans on specific tasks. Human managers must learn to validate, redirect, and exercise judgement about what AI produces, not simply execute well-defined tasks.
Krivkovich frames it as a new kind of managerial competency:
“We will need to learn how to validate, provide the right judgement, redirect, work iteratively, and test and learn.”
This is a significant departure from traditional Indian corporate culture, where precision and reliable execution of defined roles have long been the prized managerial virtues.
The new expectation is experimentation with real accountability for what worked and what did not.
The hiring dilemma raised sharply in the McKinsey research through the CHRO lens is already being felt in India.
Campus recruiters report that candidates are using AI to predict what assessment tools will select for, producing AI-optimised CVs and answers.
Several leading Indian conglomerates and consulting firms have already moved toward live case assessments, in-person problem-solving sessions, and real-time judgement tests precisely because the documented credential has become gameable in ways the in-person interaction has not.
For Brydgework clients in leadership hiring and organisational design, this is actionable today: redesign assessment frameworks so that what is being tested cannot be outsourced to a language model.
At the team level, managing human-agent hybrid teams requires rethinking every familiar management frame: KPIs, productivity benchmarks, workload distribution, and quality review cycles.
If an AI agent can generate 5,000 regulatory compliance reports overnight—a scenario already approaching reality for India’s banking regulators—the bottleneck immediately shifts to human review capacity.
As Madgavkar observes:
“We’ll be in a transition period. There’s going to be a lot of flux.”
Good managers in this era will need resilience, creative appetite, and comfort with sustained ambiguity.
The education question in the research is one of the most consequential for India.
With over a million engineering graduates annually, a vast NEET-qualified medical workforce, and an ambitious National Education Policy promising skills-based learning, India has a structural opportunity and a structural risk.
The risk:
Five-year specialised programmes in narrow technical domains may produce graduates whose primary skill is already being automated before they enter the workforce.
The five-year BTech in a single narrow engineering stream, the CA certification anchored in manual compliance work, and the medical transcription certification–all of these face the same pressure.
The opportunity:
India’s SWAYAM, NPTEL, and emerging National Skills Qualification Framework already have the infrastructure to deliver AI-tailored, just-in-time reskilling at a population scale in a way no other country can match.
The missing layer is personalisation: AI tutoring systems that ingest a learner’s specific skill gaps and deliver targeted micro-credentials at the moment of need while they are doing the work.
The McKinsey research argues persuasively that AI itself enables better reskilling: personalised to the individual, inserted at the point of work, learning from performance data in real time.
India’s enormous MOOC infrastructure—already serving over 30 million learners on SWAYAM alone—is the scaffold. The missing layer is precisely this kind of intelligent personalisation.
“There’s going to be a lot of flux. Part of being a good manager will be having an appetite for that, being resilient through it, and being able to be invested and creative through that process.”
— Anu Madgavkar, MGI Partner, McKinsey & Company
The McKinsey research is unambiguous on the question of pace: moving too slowly is the greater risk.
“Speed is a strategy in and of itself,” Krivkovich states. “By the time you get enough clarity to know definitively where to go, you’ll be far behind.”
For India, this has a particular resonance.
The country arrived late to industrial manufacturing and spent decades trying to leapfrog. It arrived early to mobile internet, and Jio’s disruptive build changed the competitive landscape across sectors in under five years.
The question is whether India’s corporate leadership, policy institutions, and education system move with the speed and coordination this moment demands.
India’s regulatory environment is a genuine variable here.
Emerging AI governance frameworks, data localisation requirements under DPDP, and sector-specific RBI and SEBI guidelines on algorithmic systems all create compliance surfaces that slow enterprise AI deployment.
The McKinsey research acknowledges this:
“Whether it’s risk management, or an understanding of ethics and compliance in the context of AI, or regulation, or public education about how to use AI responsibly, we’re going to have to move faster.”
For India’s policymakers, this is a call to build enabling guardrails, not restraining walls.
Ten years from now, the measure of success—as Krivkovich describes it—is whether AI has augmented and unleashed human potential rather than replaced it.
Whether we have moved through the uncertainty and the fear curve to a place where AI is simply part of how we work and live, like the way the smartphone and the internet are today.
For Madgavkar, the defining test is democratisation:
“Has everyone really felt the benefits? Have we seen transformational effects and wider access, and better quality? That would be real success.”
For India, this aspiration is not abstract.
A country where the top 10% of earners have disproportionate access to quality education, healthcare, and financial services has a specific opportunity here.
AI, deployed with intent and equity, could be the most powerful tool for inclusion this economy has ever seen.
Or—without that intent—it could deepen every existing divide.
Which version of that future emerges will depend almost entirely on choices made by leaders, policymakers, and organisations in the next three to five years.
The rise of the human–AI workforce is not a future scenario. It is the present reality.
The question is not whether to engage with it, but how boldly and how wisely.
1. Automate the volume and elevate the judgement.
High-volume processing (KYC, compliance, customer onboarding) is AI’s domain. Reserve human capacity for decisions that require context, relationship, and accountability.
2. Place coordinated bets, not isolated pilots
Value lives in cross-functional reimagination—supply chain + customer service + operations—not in standalone chatbot deployments.
3. Redesign hiring for the AI-augmented candidate
Live assessments, real-time problem-solving, and judgement tests matter more than credentials that can be AI-optimised.
4. Build the T-shape at scale
Every employee needs horizontal AI fluency. A few teams need deep vertical transformation capability. Both are necessary; neither alone is sufficient.
5. Speed as strategy
Waiting for regulatory, technological, or market clarity before deploying is a losing position. The cost of moving fast is recoverable. The cost of falling behind may not be.
Reference: This article is adapted for an Indian business audience from “The Rise of the Human–AI Workforce,” McKinsey & Company, April 2026.
February 13, 2026 | Hemant Kumar
Structural change in Indian banking’s distribution architecture is already an accomplished fact, thanks to the large-scale adoption of digital technologies.…
Read More
December 24, 2024 | Pranav Garg
Zoho Workplace is an enterprise email and collaboration suite that seamlessly integrates multiple applications like documents, spreadsheets, presentations, file manager,…
Read More
January 22, 2025 | Team Brydgework
The Union Government declared recently that it is working on the credit guarantee scheme for the MSME sector that was…
Read More