Four and a half feet tall, dressed in navy blue overalls and sneakers, Chandni appears almost dwarfed by the towering equipment and its blinking interface.
She moves confidently across the polished concrete floor of Wipro’s newest factory on Jaipur’s outskirts, working from a platform built just for her, elevated to meet her height.
“Earlier, some would just say she’s too short for this,” says Ranganath M.S., the Jaipur plant head of Wipro Hydraulics, watching her quietly from across the floor. “But we didn’t ask her to adjust to the machine. We adjusted the machine to her.
It’s a subtle design decision, barely noticeable unless you’re looking for it. But at this facility—one of Wipro’s most ambitious experiments in automation and inclusion—these quiet details form the core philosophy.
The Jaipur plant, which has been operational since early 2024, produces over 1,000 hydraulic cylinders daily for its largest customer, JCB. These sleek, precision-crafted cylinders power the arms of backhoe loaders and earthmovers across the country, where even a slight quality slip can dramatically ripple through construction projects.
Mistakes here are not an option.
And yet, the place hums without drama. Robotic arms swivel methodically, conveyors glide under high ceilings and human operators—many of them young women—oversee the precise movement of steel rods and components. The work is physical, but no one hauls heavy metal on their shoulders. “We’ve gone from muscle to brain,” Ranganath explains. “Most handling is automated. Human work now means judgment, observation, thinking.”
Sensors enforce that thinking. If an operator skips cleaning a welding nozzle, the entire system comes to a halt. “You can’t proceed without compliance,” says Reena Praharaj, the plant’s human resources lead. “Mistakes aren’t corrected after they happen—they’re prevented before they begin.”
View Full Image
Overhead, soft green lights blink approval with each cycle, monitored by HERCA, the factory’s human-error prevention system.
For Chandni and many others, this factory represents more than just a job. Older shop floors often meant years of rote repetition and exhaustion. Here, every operator is a diploma-holding engineer, rotating through roles, gaining mastery and preparing to rise. Many, like Chandni, are in their first job—away from home for the first time, sending money back to families who doubted they could manage at all.
Ranganath shrugs when asked how this became possible. “You start with a simple principle,” he says quietly. “If she were my daughter, what would I want her workplace to look like?”
In Jaipur, that means the floor itself has risen to meet the worker.
And Chandni isn’t alone. Priyanshu Pandey, from Mirzapur in Uttar Pradesh, faced resistance too. Her family expected her to follow the conventional path: arts in school, then a job at her mother’s beauty parlour. Instead, she pursued mechanical engineering, cycling 13 kilometres daily just to study. “People said, CS (Computer Science), IT are better for girls. You won’t manage mechanical,” Priyanshu recalls. “But I thought, if boys can, why can’t I?”
Today, Priyanshu proudly operates a CNC machine at Wipro Jaipur, reassuring girls from small towns that mechanical engineering isn’t beyond them. “AI and sensors aren’t magic,” she says. “They just need understanding.”
Women like Chandni and Priyanshu now constitute 36% of the workforce at the Jaipur plant.
Thinking factories
As global supply chains realign and India makes its bid to become the world’s next factory floor, a sobering gap remains: manufacturing contributes just 13% to the country’s GDP, far behind East Asia’s 25–30%. Billions in incentives have moved the needle slowly. The real breakthroughs may not come from policy, but from places like Wipro’s hydraulics division—its oldest, least flashy business—where artificial intelligence (AI) is quietly redefining how factories work.
At Wipro Infrastructure, AI now predicts machine failures with over 90% accuracy. Tool breakage has dropped to zero on over 100 machines. Defect rates have dropped. The shift isn’t just technical—it’s philosophical: from reaction to prevention, from instinct to insight.
“AI isn’t about headcount reduction,” says CEO Pratik Kumar. “It’s about amplifying human judgement—and scaling it.” He sees AI as the lever to grow Wipro’s non-IT verticals into multi-billion-dollar businesses.
View Full Image
For a company turning 80, the arc from Amalner’s oilseed presses to sensor-wired shop floors is more than evolution—it’s strategy. In an era where India’s industrial future depends not just on cost but also on competence, Wipro is testing a new model: factories that think, workers who adapt, and automation built not to replace humans but to help them excel.
And the stakes go well beyond Wipro. Globally, discrete manufacturing loses nearly $300 billion a year in inefficiencies. Even top plants run at just 65–70% productivity. But as Ranjit Date, the engineer behind Wipro’s Linecraft AI, notes: “With the right models, that number can reach 90%. That’s less capex, less waste—an entirely different way of valuing manufacturing.”Wipro Infrastructure acquired Linecraft AI, a predictive intelligence company, in 2022.
The Legend of Ajay Bihari Lal
Jugal Prasad still remembers the sound.
“It was ’97 or ’98. I was 24, fresh out of IIT Kanpur, walking the Tata Motors shop floor with my supervisor, Ajay Bihari Lal,” Prasad, vice president at Wipro Infrastructure, recalls.
Mid-step, Lal paused suddenly. “He just stopped and said, ‘Something’s wrong with this machine.’ No dashboard, no sensor—he’d caught a subtle noise.”
The culprit, it turned out, was a bearing failure in a German Giesen gear-cutting machine—a breakdown Ajay sensed with nothing more than his ear and experience.
“That was always in my mind,” Jugal says. “How do we remove that skill? Because that skill is no longer available.” In the world Jugal helps shape at Wipro today, the ear has been replaced by sensors, the intuition by algorithms. Machines carry hundreds of sensors that monitor everything from noise and vibration to temperature and friction in real-time. “If something is not going well,” he explains, “the temperature of the components or machine will go up… more friction means more heat… and those are the signals that come out before it fails.”
View Full Image
Across critical machines, 72 sensors—tracking vibration, temperature, and noise—collect tens of thousands of data points daily, monitoring machine health in real time. “What used to depend on someone’s instinct now depends on predictive models,” Jugal says. With 95% accuracy, Wipro’s algorithms forecast machine failures hours in advance, cutting downtime to one-tenth of its earlier levels.
Then there’s HERCA—Human Error Root Cause Analysis—designed to prevent mistakes proactively. Take the application of loctite glue, a crucial for parts to hold securely. “If an operator skips loctite, the part may pass initial tests but fail in the field after 1,000 hours,” Jugal says. At Wipro, sensors ensure that the glue bottle is lifted before proceeding. “Unless I lift the loctite bottle, I am not able to move to the next operation. The cylinder gets locked.”
“We are trying to maximize our output from the process… with the least cost,” he adds.
The aim is deceptively simple: machines that run longer, reject less and produce more with fewer skilled hands. “So if in a day we’ve planned for 22 and a half hours, and the machine is running for 22 and a half hours efficiently, then I’ve maximized my output.”
But behind that efficiency is a world of real-time prediction. On one bottleneck machine, for instance—a line that could halt the entire plant if it fails—Wipro has installed six sensor-loaded stations, each with 12 sensors, streaming 72 data points. “We take the average of 10 minutes of those 72,” he says, “and on a real-time basis, it gets analysed. It tells us what is the likelihood of a breakdown in the next 12 hours. We are almost 95% accurate in that.”
Across critical machines, 72 sensors—tracking vibration, temperature, and noise—collect tens of thousands of data points daily, monitoring machine health in real-time.
Before this system, maintenance was either reactive—fix it after it breaks—or scheduled blindly every few months. “Now it’s condition-based,” Jugal says. “We fix things before they fail and avoid all the consequential damage.”
A failed bearing today, he explains, is like a flat tyre: “You don’t just change the tyre, the rim gets damaged, the alignment goes for a toss.”
This kind of precision isn’t replacing people—it’s empowering them.
The rise of the invisible engineer
Maitrik Siddhapura, a machine vision engineer at Wipro PARI,an automation company in Pune, represents a wave of new AI jobs emerging at the intersection of sensors and software. His role—designing systems that catch micron-level defects on the line—didn’t even exist five years ago.
View Full Image
His childhood was shaped by visits to over 50 factories—tile presses, forging plants, machine shops—thanks to his family’s sheet metal business. “We made CNC press brakes and power presses,” he says. “I’d tag along just to see how things were built.”
That early curiosity turned into an obsession. In college, Maitrik taught himself machine vision—how to make machines see—and today, as a vision engineer, he’s part of a rising generation that is quietly redrawing the shop floor. His job begins before any camera is mounted: choosing lenses, lighting, specs, and then training the system to catch flaws a human might miss. A barcode printed at the wrong angle. A missing pill in a sealed blister pack. A lipstick tube with a micron-level scratch.
On one project, misfit parts threatened a production delay, so Maitrik improvised with components from another line. On another occasion, he had to rethink 2D scanning altogether. “We were reading barcodes on reels that varied from 900 to 1400 ml. The curvature threw off the angle. That’s when you realize: vision isn’t just optics. It’s logic.”
His team has worked across FMCG, pharma, and automotive lines, where speed and precision aren’t perks—they’re prerequisites. “You’ve never seen an empty cavity in a medicine strip,” he says, “because a vision system won’t let it pass.”
But what drives him isn’t just the tech. In college, he built a machine learning model to sort waste on a conveyor belt—separating plastic, paper, and glass. “In India, humans still do this by hand. That’s not just inefficient, it’s inhuman. We can do better.”
That belief led him to donate a 3D vision system to his alma mater. “Let’s make this field less niche. If I can learn it, so can others.”
Maitrik estimates the demand for machine vision will double in five years. But the deeper shift is already here. “These systems don’t just see now. They judge. They adapt. That’s exciting—but it also means we need to get it right.”
He smiles, remembering how it all began—with one cold email, one line: “I’m crazy about robotics. I’ll do anything to join PARI.” Someone replied. He never looked back.
Maitrik estimates the demand for machine vision will double in five years. But the deeper shift is already here. These systems don’t just see now. They judge.
In Chennai, another young engineer named Aditya oversees a predictive dashboard, crunching sensor data into actionable insights. He points to 29 January: “We flagged a spindle 16 hours before it failed. Zero downtime.” Out of the 42 recent potential breakdowns, Aditya’s AI system correctly predicted 38—over 90% accuracy. “We’re no longer reacting,” Aditya explains calmly. “We’re anticipating.”
What Lal once did by instinct, engineers like Maitrik and Aditya now achieve through carefully taught algorithms and thoughtfully deployed sensors. Intuition hasn’t disappeared—it’s just become data-driven.
Listening to machines
Ranjit Date clearly remembers the first time he saw a CNC machine. It was the early 1980s, and Date was a young intern at Kirloskar’s factory in Pune. “Computers were controlling machines,” he says, smiling gently at the memory. It was silent, precise, and magical.”
At that time, India had only four industrial robots—“Two at Maruti, one at Bajaj, one at Tata Motors,” Date recalls fondly, as though naming old friends. The idea that these machines would soon populate India’s factories seemed absurd to most, but Date was captivated by the possibility.
In 1990, from a small garage in Pune, Date co-founded PARI (Precision Automation and Robotics India) with
View Full Image
an audacious goal: to bring world-class automation to Indian manufacturing. Friends laughed at the poster on his wall listing the “Top 10 global automation companies”—with PARI optimistically penciled in. “Azim Premji once said, ‘If people aren’t laughing at your goals, you’ve set them too low,’” Date says quietly. “I never forgot that.”
Wipro Infrastructure acquired PARI in 2021.
In March 2025, after 35 years, Ranjit Date retired as CEO of Wipro PARI, passing the leadership to Vipul Tandon.
Another of Date’s startups, Linecraft AI, was also acquired by Wipro. Its predictive intelligence now powers Wipro factories, dramatically boosting productivity—from an industry standard of 65% efficiency to nearly 90%. The impact, Date says, isn’t just efficiency: it’s economic transformation: less downtime, fewer lost hours, less capital expenditure.
“We aren’t just running factories anymore. We’re listening to them.”
For Sitaram Ganeshan, president of Wipro Hydraulics, the real transformation isn’t in the toolset—it’s in the questions. “AI isn’t about solving grand problems,” he says. “It’s about asking: can this be done better?”
From engineering to HR, he wants every team to start there. The competitive edge, he insists, doesn’t lie in technology alone but in a culture willing to ask—again and again.
He’s begun posing new questions: Where is the robot wasting time? Can value addition be measured in seconds? Could we run a floor where no one flips the lights on?
“If ERP (enterprise resource planning) is widespread but underused,” he says, “AI could go the same way—unless we build a mindset around it.”
Dark, but not empty
Standing on Wipro’s shop floor in Chennai, Jugal surveys the synchronized ballet of robotic arms and sensor-guided conveyors, then quietly says, “We’re close to a truly dark factory. Lights out, fully automated.” He pauses thoughtfully. “But it’s not science fiction anymore. Maybe five years, not 10.”
In manufacturing circles, “dark factory” evokes visions of stark automation—a plant without lights or human presence, fully autonomous and efficient. It’s often viewed with anxiety as a stark endpoint of job displacement. But at Wipro’s plants in Jaipur, Chennai, and Pune, the concept feels closer to thoughtful evolution than dystopian replacement.
Jugal recalls an early Tata Motors memory, shaving gears for the Tata Sumo. “We’d load 40 gears, and three hours later, finished pieces emerged. Imagine automating that fully—no intervention needed, 24 hours straight.” He smiles faintly. “In Jaipur, we’re almost there. Raw rods go in, finished components come out every minute. Continuous feeding is all that’s left to perfect.”
Across the factories, critical functions—such as tool-wear detection, lubrication checks, and even subtle vibration analyses—already happen automatically. But this isn’t about sidelining people, insists Jugal: “Machines don’t tire, they don’t skip tea breaks. Humans are still crucial for innovation, troubleshooting, and adaptation.”
View Full Image
Ranjit Date, whose Linecraft technology underpins Wipro’s shift to predictive, sensor-driven efficiency, frames it starkly: “Human costs rise, machine costs fall. But AI isn’t about blindly replacing people—it’s about erasing drudgery, waste, inefficiency.”
If operators no longer carry heavy rods on their shoulders, he asks, “How can they contribute more meaningfully?”
In Jaipur, the factory’s thoughtful design—platforms adjusted to shorter workers, along with meticulous safety protocols—demonstrates that humans remain central, with their roles elevated, not erased.
The dark factory, then, isn’t the end of human labour. It’s the beginning of a different kind of work—smarter, safer, more intentional. What feels revolutionary today, Jugal insists, is tomorrow’s reality. “We’re already there, station-by-station. Once we connect everything, the lights might go off—but the thinking won’t.”
When the machine doesn’t know why
As AI quietly transforms how Indian factories run, one of the deeper risks lies not in failure but in not understanding why failure happens.
“Manufacturing is often seen as a closed system,” says Jibu Elias, an AI policy expert. “Because it’s not consumer-facing, we haven’t had a serious conversation about what responsible AI means in this space.”
At first glance, the risks may seem limited. Wipro’s industrial AI systems don’t process personal data. Their predictive maintenance algorithms analyze sensor signals—such as vibration, heat, and torque—not faces or speech. But, as Jibu points out, the stakes are still real.
“Explainability is critical,” he says. “When something goes wrong on the shop floor, it’s not enough for the model to say ‘halt the machine.’ You need to know why it’s halting. What’s the root cause? Can an engineer intervene—or will the system keep failing in the dark?”
He shared the story of a failed AI experiment in New Zealand’s dairy industry. An automated system trained to process milk into powder kept producing defective output. The algorithms were retrained, and the sensor streams recalibrated—still no luck. “Eventually, a dairy scientist walked in, checked the pipe, and found a fungal infection. That was it,” Jibu says. “You still need domain expertise. AI can’t smell fungus.”
This, he stresses, is where Wipro and others must tread carefully. “You can’t design out the human. You have to design for humans.”
One subtle but rising concern, he adds, is the issue of algorithmic control over labour, where systems begin to dictate work at such granularity that autonomy disappears. In India, where labour is still cheap and compliance thin, the risk isn’t just displacement. It’s dehumanization.
That’s why, Jibu argues, the building blocks for responsible industrial AI must include more than just model accuracy or uptime. “You need explainability, yes. But also checks on how the system interacts with labour. Where is the dignity? Who has override access? Is this system making work safer—or just more monitored?”
He adds: “The question isn’t whether AI will come to the factory. It’s who it listens to when it gets there.”
Amalner to Jaipur
About 80 years ago, in Amalner’s modest stone factory, there were no sensors or algorithms—just groundnuts drying under monsoon skies and quiet men measuring productivity by instinct. Azim Premji once calculated risk in fistfuls of lost groundnuts; today, in Jaipur, sensors pause entire assembly lines if Chandni, elevated carefully on her custom-built platform, misses even a single step.
Wipro’s core values remain constant: discipline, frugality, and quiet respect for human dignity—now amplified by industrial AI. Yet, as factories across India embrace automation, urgent new questions arise: the very concerns highlighted by experts like Jibu Elias, about algorithmic oversight, labour privacy, and the explainability of machines that increasingly guide human hands.
In Jaipur, Chandni finishes her shift, sends part of her paycheck home, and quietly expands the possibilities for daughters everywhere. The factory of the future isn’t just built from steel, sensors, and AI—it’s shaped by how thoughtfully we balance human potential and machine precision.
