“My own rule would be this: never ask a human to rubber stamp an AI recommendation; ask them to own the consequence of accepting or rejecting it. That changes the quality of decision making immediately,” observes D K Rai, CEO India and VP Global Business Expansion, Smartlog Corp. His perspective reframes expansion in the AI era as a leadership challenge—where growth is measured not only by speed and scale, but by the responsibility leaders take for the outcomes their choices create.
AI will not replace supply chain judgement. It will expose whether we ever had good judgement to begin with. That, to me, is the real shift underway. In India, we often talk about AI in supply chains as if the destination is a “lights-out”, fully autonomous system. I think that is the wrong metaphor for our market. India is too dynamic, too fragmented, too infrastructure-sensitive, and frankly too human for that fantasy. Our roads, vendors, demand signals, regulatory realities and channel behaviour still refuse to behave like a clean spreadsheet. So the future here is not human-less supply chains. It is human judgment operating at a higher altitude. Indian enterprises are already moving quickly on AI adoption—including in strategy, operations and supply chain—but capability depth still lags, which is exactly why human quality will matter even more in the next phase.
How will human decision-making change as AI scales?
I do not believe humans will become less important. I believe routine human intervention will become less important. The planner of the future will not be valued for producing a forecast file, chasing updates on WhatsApp, or sitting in review meetings explaining why yesterday’s numbers changed. AI will increasingly do the sensing, the pattern detection, the recommendation and, in some cases, even the execution. The human role will move upward — from processing decisions to framing them, challenging them, and carrying the commercial and ethical responsibility for them. That is a profound shift. It means the real premium will be on judgment under ambiguity, not on mechanical coordination. That is also where the best AI performers are separating themselves: they are not merely automating old work; they are redesigning workflows and defining where human validation is required.
In practical terms, I see the human role becoming more like an air-traffic controller than a ticketing clerk. Less touching every transaction, more supervising the system, interpreting exceptions, and deciding when the algorithm is technically right but contextually wrong. In India especially, context still matters enormously: a port delay is not just a delay, a festival spike is not just a spike, and a supplier commitment is often as much about relationship temperature as contractual SLA. That is why I say AI will compress routine judgment, but it will increase the value of seasoned judgment.
“I learned this the hard way during the COVID crisis, when the system gave us one answer, but the ground reality demanded another.”
Which traditional roles are losing relevance, and what new roles are becoming critical?
The roles losing relevance are not entire job families overnight; it is the parts of jobs built on low-value repetition. The classic expeditor who spends the day chasing status updates, the analyst who manually reconciles spreadsheets from five systems, the planner whose identity is tied to monthly file-making rather than decision quality, and the coordinator who adds no insight beyond moving information from one inbox to another — these roles will steadily lose relevance.
What will grow instead are hybrid roles. We will need supply chain professionals who can act as decision architects, not just process owners. We will need people who understand data, but also know operations well enough to spot when the data is lying. We will need AI product owners inside supply chains, scenario designers, model-governance leaders, exception managers, control-tower orchestrators, and translators who can convert business ambiguity into machine logic. McKinsey’s framing is useful here: gen AI becomes a “team member” that must be trained, given business context and improved through feedback. That means new human value lies in teaching the machine, supervising it and redesigning work around it.
If I were being blunt, I would say this: in the old supply chain, being busy was enough. In the new supply chain, you have to be useful. AI will be ruthless about that distinction. “Early in my career, I was rewarded for speed of response. Today, I value quality of intervention far more — knowing when to step in, when to let the system run, and when to escalate.”
What skills must the next-generation supply chain professional develop?
The first is systems thinking. A lot of people can optimize a node. Very few can understand what that optimization breaks somewhere else. In an algorithm-driven supply chain, local efficiency can create network fragility. So the future professional must think in terms of trade-offs, feedback loops and second-order effects.
The second is data literacy without data arrogance. People do not need to become coders to remain relevant, but they do need to understand what data is being used, what is missing, what assumptions are embedded in the model, and where bias or blind spots may sit. The fastest-growing skills globally now include AI and big data, technological literacy, creative thinking, resilience, curiosity and analytical thinking — and that combination matters. The next supply chain leader must be both numerate and adaptive.
The third is commercial judgment. AI can tell you what is probable. It cannot fully tell you what is worth doing. That still requires an understanding of customer strategy, working capital, market share, supplier relationships and risk appetite.
The fourth is narrative ability. This is underrated. In the future, the person who wins will not be the one who has the most dashboards, but the one who can explain what the dashboard means, what decision is required, and what trade-off the business is actually making. And the fifth is the courage to challenge clean-looking output. In my view, one of the most important future skills is not prompt engineering. It is respectful scepticism.
“I tell young teams that in the AI era, your career will not be protected by knowing more screens than the next person. It will be protected by asking better questions than the next person.”
How do we bridge the gap between current talent and AI-driven capability needs?
The biggest mistake companies make is treating this as a training problem alone. It is not. It is a work redesign problem first, and a talent problem second. If an organisation simply runs AI workshops but leaves the underlying workflow unchanged, people will go back to old habits. The more effective approach is to redesign a few high-value decisions end-to-end: forecasting, replenishment, supplier risk, transport planning, inventory deployment. Then train people inside those redesigned workflows. High-performing AI organisations do exactly this — they redesign work, define validation checkpoints, and back it with leadership ownership, not just tech enthusiasm.
In India, bridging the gap requires a very practical model. First, create mixed teams of operators, planners, data people and business leaders. Second, use apprenticeships more than classrooms. Third, rotate high-potential talent through plant, procurement, logistics and analytics roles so they become bilingual in operations and algorithms. Fourth, reward adoption and decision quality, not only course completion. Deloitte’s India data is telling: Indian firms are scaling AI quickly, but expertise depth is still low, so the imperative is not just to deploy tools but to build specialist and managerial capability around them.
India also has a broader workforce challenge and opportunity here. Deloitte and NASSCOM have noted that AI talent demand in India is set to more than double across 2022–27, while many workers already see AI skills as career-enhancing. That tells me the answer is not to wait for the perfect talent market. It is to convert current supply chain talent into AI-capable operators through structured upskilling, live use cases and cross-functional exposure.
“The companies that win won’t necessarily hire the most data scientists. They’ll be the ones that best reskill their toughest planners, buyers and plant managers into human supervisors of machine-led decisions.”
How are organisational structures and leadership models changing?
The old supply chain organisation was designed around process control. The new one is being designed around decision velocity. That means the hero model is changing. The celebrated leader of the past was often the person who could personally unblock every crisis. The leader of the future will be the one who builds a system where fewer crises require heroics in the first place. So leadership is shifting from command-and-control toward orchestration, from vertical silos toward cross-functional pods, and from static reviews toward dynamic control-tower environments.
The language matters here. “Execution” suggests people moving tasks through a chain. “Orchestration” suggests humans, machines, partners and platforms working as an adaptive network. The World Economic Forum describes this evolution as a journey from digital to adaptive to autonomous supply chains, with talent, governance and cross-functional collaboration as critical enablers. IBM makes the same point differently: as AI agents act within defined constraints, human teams move toward setting objectives, handling exceptions and overseeing outcomes.
In India, I expect the strongest organisations to build what I would call a federated intelligence model: central standards for data, governance and architecture, but decision rights distributed close to the market, plant, customer or supplier. That suits our complexity far better than importing a rigid global operating template.
“In my own experience, some of the worst supply chain decisions were not made because people lacked data; they were made because the data was trapped in silos and nobody owned the full outcome.”
In high-stakes situations, how should organisations balance AI recommendations with human judgement?
My view is simple: the higher the stakes, the more explicit the human role must become. I do not believe in blind trust in AI, and I also do not believe in romanticising human intuition. Both are dangerous when left unchecked. The answer is a risk-tiered model. For routine, high-volume, low-consequence decisions, let the machine decide within thresholds. For medium-impact decisions, let AI recommend and humans approve. For high-stakes situations — major allocation choices, crisis response, safety risks, compliance exposure, strategic customer service trade-offs — require a visible human decision owner, with model confidence, assumptions and escalation triggers made transparent.
What will differentiate supply chains in an AI-driven future?
Not technology access. That will commoditise faster than most people think.
For India, I think the winners will be the companies that combine three things unusually well:
- The frugality to scale AI pragmatically rather than fashionably;
- The operational realism to build for messy ground conditions rather than idealised ones;
- The talent discipline to turn current managers into orchestrators.
India has strong momentum on enterprise AI adoption, but many firms are still early in maturity and continue to face barriers in quantifying value, changing behaviours and building ecosystem readiness. That is why talent capability will be the real accelerator: it converts AI from a pilot into a compounding advantage.
For years, supply chains rewarded endurance — the people who could tolerate complexity, chase updates, and keep the machine moving. The next era will reward something else: judgment. AI will handle more of the motion. Humans will be judged by the quality of their intervention. That is why I don’t think the future belongs to autonomous supply chains in the purest sense, especially not in India. I think it belongs to intelligently human supply chains — systems where machines scale speed, and people scale wisdom.
Categories
