Why 80% of the World's Workforce Never Got an AI Product, and What Ankush Jagga Is Doing About It

A few months ago, India's Prime Minister Narendra Modi put on a pair of AI powered smart glasses, built in India, at the Indian AI Summit. The moment made headlines, but honestly, it was never really about the glasses. It was a signal. AI is leaving our laptops and our screens and finally stepping into the real world.

Here's the thing though. Most AI products today are built for people who sit at desks and do digital work. But 80% of the world's workforce doesn't sit at a desk at all. They're out fixing your air conditioner, servicing your car, picking your orders, climbing up windmills, and standing on airport tarmacs. Most of the time, they're relying on manuals, on memory, and on whoever senior person happens to be free on the phone.

Ankush Jagga has been building businesses since 2003. That's 23 years, three startups, and two exits. Now he's working on what he believes is the most important one yet: AI that works right where the job is happening, for the frontline workers the tech world forgot to build for.

He's the co-founder and CEO of Unfold XR, and this conversation gets into why smart glasses might finally have their moment, what it actually takes to build software across a fragmented hardware landscape, and why he thinks India's scale could change the entire category.

Key Takeaways: Building AI for the 80% Who Never Sit at a Desk

The Core Problem:

• 80% of the global workforce is frontline, not desk-based, and almost no AI product has been built with them as the primary user
• Frontline workers today rely on memory, manuals, and phone calls to supervisors, with no real-time validation of the work they're doing
• Unfold XR pairs an AI expert and a human expert with the worker at the same time, both seeing exactly what the worker sees

Hardware Reality Check:

• High-end industrial smart glasses cost $2,000 to $5,000, which Ankush realized early on "can't scale at this price"
• Smart glasses today fall into four broad categories: no-screen consumer glasses ($50 to $100, like ones out of China), industrial-grade glasses built for hazardous environments, and glasses built for 3D or 6DoF experiences
• Unfold XR is deliberately device-agnostic, building the same experience for a $100 smartphone and a $5,000 industrial headset, because "we're not married to what device you're doing, we're married to improving that human life"

Proof Points and What's Next:

• Unfold XR has run a proof of concept and pilot at a major commercial airport, digitizing the paper checklist ground staff use for aircraft inspections, with around 50 people already on the pilot
• The company measures pilots on two KPIs: time and accuracy, both of which Ankush says are "hitting very well"
• Ankush's vision is a million-plus users, in a category where he says no competitor has even crossed 100,000

Q: For a layman, what does Unfold XR actually do?

Ankush Jagga: Okay, imagine fundamentally we're helping people to do their job better. Imagine I'm a worker. I go to my site. The old word was I go to a place of work. Somebody tells me what is to be done, I do it. If I have a problem, I'll call a supervisor. Supervisor is no help, I get help, I do not get help, I'm on borrowed help. Whatever I'm doing, nobody's validating it.

Now go to the new word. Somebody's wearing glasses like these. The vision computing is telling me where I am, validating the job I'm doing, telling me what jobs I have to do step by step. And during this entire process, if I have a problem, I have two experts with me. I have the AI expert and I have the human expert, and the expert now gets to see the exact same point of view that I see. If I'm pointing at a machine, the expert also gets to know what part I'm pointing at.

We're trying to build a B2B experience on how smart glasses can be consumed well for frontline workers. And when I say frontline workers, I don't mean just the workers. The spectrum is from a doctor to a construction site worker. A doctor is also not on a desk.

🔥 ChaiNet's Hot Take: Every AI product roadmap talks about "the future of work" while quietly assuming that work happens at a keyboard. Unfold XR is betting the real opportunity was never in the inbox, it was 100 feet up a windmill.

Q: Give us a real example, like the windmill technician story you've mentioned.

Ankush Jagga: Think I'm a windmill technician whose job is to service wind turbines, 100 feet above the ground. Once every 20 or 30 days I have to service it. Not every decision is straight line. I'll see some reading there and that reading tells me either to replace something or service something. You expect me to remember all of this as a technician? I don't remember everything. Today as humans, we have evolved a lot.

What if the vision computing is looking at it and telling me, "look, I see a heat reading of 60 degrees, you should be changing this belt," or "I see a combination of these three inputs, based on which you should be performing a detailed cleaning." In the old world, how would this person have taken support? They would have identified the issue, come down, made a call. You don't expect me to pull out my cell phone and make a call up there. All of this assistance is there now. And still, if AI isn't able to help me out, I have a video call facility within these glasses, and the expert can see my point of view.

🔥 ChaiNet's Hot Take: A windmill technician calling a supervisor from 100 feet up isn't a workflow problem, it's a safety hazard waiting to happen. Unfold XR isn't just saving time here, it's removing the moment where someone has to choose between doing the job safely and doing it fast.

Q: Where are you in your journey with the proof of concept and pilots?

Ankush Jagga: It's a very interesting journey. We have done a proof of concept, we've done showcases at multiple areas, including for a major commercial airport, one that all of us would have gone to at some point. We showcased it to DGCA over there. So the product is there, the product market fit is there.

The word tried solving these cases with very high-end tech, device-specific use cases. That's an expensive piece of component, anywhere from $2,000 to $5,000. Very early we realized this can't scale at this price. So what we started building is fundamentally instructional, irrespective of what device category a user is on. Irrespective if the user is using a $100 device or a $5,000 device, we're trying to build it all across. Tech is there, tech has been built. We're now adapting it to different hardware devices, and at the same time opening up more use cases. Repair is a use case, inspection is a use case. That paper-based inspection before you get your car delivered, checking if the paint is right, the electricals are working, we're moving that to a smart glass identifying "this is chipped a little" and flagging it.

🔥 ChaiNet's Hot Take: Most hardware-first startups get seduced by the most expensive, most capable device on the market. Unfold XR did the opposite: it looked at the price tag first and decided the real product was the layer that works on anything.

Q: What's the range of smart glasses that can actually support this?

Ankush Jagga: Broadly there are three or four categories of smart glasses. One, you have the Meta Ray-Bans. I'm so happy Meta solved a big education problem for us. When I was incepting this business, my parents asked what a smart glass even was. Today, thanks to Meta, that education problem is solved.

There are smart glasses available in China for $50 to $100 with no screen, just camera, mic, and speaker, good battery, 60 to 80 grams. What I'm wearing right now is almost 40 to 45 grams. Then there's the other extreme, industry-approved smart glasses for environments like an oil rig where there are flammables. These come with better battery backup, infrared inside so they can identify heat, and they're built to survive rain and bad temperatures. In between, there are glasses built around 3D and 6DoF, six degrees of freedom, experiences.

Google Glass launched years back and scrapped the project because they were ahead of time. They've relaunched it now. Apple is getting into smart glasses. The tech giants are saying smartphones are going to go off, and we'll have some device continuously telling us all the information we currently see on our smartphone. It can be lenses tomorrow. With Unfold XR, we're not married to smart glasses, we're married to augmenting the human, to helping the human do better. It can be a neuro device tomorrow, a brain chip, whatever it turns out to be. That's what we're married to.

🔥 ChaiNet's Hot Take: Betting on a specific device is a bet on a company you don't control. Betting on the human doing the job better is a bet nobody can obsolete.

Q: With so many different smart glasses running different platforms, do you have to rebuild the whole system for every new device?

Ankush Jagga: You're talking about the same thing my team tells me every day, and I look at it as an opportunity. As a founder, all problems are opportunities, because they're problems for somebody else too. My only edge against competitors is that they will hit the wall and give up, and I will hit the wall and cross it anyway.

You're right, every smart glass runs on a different platform. Twenty to thirty percent of the job has to be redone when you move from one smart glass experience to another. It's reducing though. Smart glasses came twenty years back and nothing happened, because we couldn't think of use cases, and the hardware was bulky. I'll give you something in my defense: Jio has jumped into this business, they've launched lenses. And we all know if Jio jumps into something, they're going to democratize the entire market, and nobody can sell that piece of hardware if there's no software on it. It's their responsibility to make software development easy on it. If it's too complex, nobody will buy it.

🔥 ChaiNet's Hot Take: When a telecom giant with hundreds of millions of users enters hardware, the smart move isn't to compete with them, it's to become the software layer they can't launch without.

Q: How are you planning to price this for the Indian market, given it's a B2B product to begin with?

Ankush Jagga: For us it's a B2B product. I was talking to somebody at the Ministry the other day, and they asked me, "Ankush, why are you only thinking B2B? Why can't you build something people can buy, like people take courses, why can't somebody use this as an assisted device to perform their job?"

Let me give you an example instead. Did you see the Bombay AI Summit showcase where Urban Company showed how a geyser mechanic learns to fix your geyser when they come to your house? They created an AI agent that has context of which house, which geyser, which company, what problem, and now gives instructions to help solve that problem for that technician on-site. We've taken it a step ahead. We're saying app, AI, and augmented reality, coupled together, that's the experience we're building.

Somebody told me, if you can make it useful, people will value it. That's the responsibility we have. With respect to pricing, globally it's very expensive, but we'll figure it out. If it's useful, you'll get the right customer set, you'll get the right scale. Today my problem isn't the pricing, it's making it useful. If we make it useful, I think pricing finds its way on its own.

🔥 ChaiNet's Hot Take: Most founders chase pricing strategy before they've earned the right to charge anything. Ankush is doing the unglamorous thing first: proving usefulness before touching the price tag.

Q: Why does India specifically give you a different kind of scale here?

Ankush Jagga: India is going to give you a very different scale than the rest of the world has given. There are so many big companies in the world, but nobody has attempted this on so many device categories and so many instruction modes, voice only, augmented environment, 3D, 6D. We're attempting it on so many different avenues. We don't want everybody to buy a $5,000 device. You have a $100 smartphone? Please use it on that. We're opening up the window much higher.

If you look at the competition, nobody has even crossed 100,000 users. Our vision is a million-plus users. We've imagined it very differently because of this range we've opened up, and I think that gives us scale, and in that scale, it's going to bring price down and give the right value to the market.

🔥 ChaiNet's Hot Take: Everyone else in this category optimized for the best possible device. Unfold XR optimized for the most possible users. In a market the size of India, that's not a compromise, it's the whole strategy.

Q: How does this compare to what a company like Urban Company is doing with its own AI-guided technicians?

Ankush Jagga: Think about why Urban Company is also doing this. They'll also lose the second service if they don't. They're trying to build an ecosystem where the technician is also dependent on it. Humans have evolved, we don't remember things the way our fathers used to. So they're ensuring their technicians are dependent on this tech, so they don't lose the revenue. Next thing, the technician might tell the customer, "don't come through Urban Company, come directly and I'll give you a better rate." They're insuring against that too. All of this is built on the thesis that humans have evolved.

🔥 ChaiNet's Hot Take: Every marketplace's biggest fear isn't a competitor, it's disintermediation by its own supply side. Smart assistive tech quietly solves that by making the technician better at the job only inside the platform's walls.

Q: What was the pilot you ran with the airport?

Ankush Jagga: I remember sleepless nights. It's still going on while we speak. Imagine DGCA rules, and doing video shoots and drone shoots at an airport. Some areas we got approvals, some areas we're still waiting on, data has to stay in India, a lot of complications.

But towards the end, what we were able to showcase: anytime you take a flight, you'd have seen ground staff with a paper, writing down certain things, a checklist they're filling. Imagine that paper going off. We were training the system on different types of grease leaks. If a grease leak happens and the technician doesn't catch it and puts a tick mark saying "no grease leaking," the camera is always on and it can identify the grease and flag it. I can still override it, but it flags it in the system. I'd be scared as a technician, honestly, but it's helping me do my job better. If I do see the problem, I press a button and it takes a picture and files it. The moment a fault is reported, our inventory and skill system automatically identifies the next best technician available to fix it.

🔥 ChaiNet's Hot Take: A checklist tick mark is only as honest as the person filling it out. Unfold XR didn't digitize the paper, it removed the option to lie on it.

Q: For something as critical as a pre-flight checklist, does this change the safety equation too?

Ankush Jagga: Absolutely, things become hard to miss. There was a learning curve initially, but we've tested it on around 50 people, and they've been performing better every time they're on it.

🔥 ChaiNet's Hot Take: In aviation, "hard to miss" isn't a nice-to-have feature, it's the entire point of the checklist. AI just made the checklist actually do its job.

Q: What are the KPIs you're tracking on a pilot like this?

Ankush Jagga: Time and accuracy, these are the two very important things. Earlier, I used to take pictures and upload them, and nobody used to look at it, no AI validation. Now if I upload a picture, AI is by default validating it. Time, and accuracy. I think we're hitting both of these KPIs very, very well.

🔥 ChaiNet's Hot Take: "Nobody used to look at it" is the quiet admission behind most manual compliance processes. The paperwork existed, the accountability didn't.

Q: Do you think this is the year augmented reality gets serious in India?

Ankush Jagga: All thanks to the consumer, B2C side of the business. Let me tell you what's happening in North America. Amazon has committed that in the next two years, everybody in North America is going to ditch the QR scanners they use, logistics in North America is going to be on smart glasses. If we can produce even 30 seconds out of a delivery timeline, with a business like Blinkit's ten-minute delivery, that could mean building 5% less dark stores in an area. That saves a lot.

If North America is adopting this, our challenges are even more, so we want to monitor things even more, we want assistance even more. And I think we've become very tech forward as a community, way more tech forward than most of our counterparts, ready to adapt and adopt new technology quickly.

🔥 ChaiNet's Hot Take: When Amazon commits to ditching QR scanners across an entire continent, that's not a trend piece, that's a market opening up a two-year runway. The founders paying attention right now are the ones who'll own it.

Q: What would you tell a first-time founder, something it took you 23 years to learn?

Ankush Jagga: Patience, perseverance, and discipline. I was 19, I'm 42 now. All this while, if I hadn't had patience, I wouldn't have been able to continue like this. You will continuously fail, always. Just stick to what you do, have patience, have faith, have belief, persistently do that, and have the discipline.

A founder is a full-time job, it's a myth when people say it's not a real job. It's actually much more than a job. If you don't have your discipline right, you can't call yourself a founder. Anything less than 24 hours a day is part time.

🔥 ChaiNet's Hot Take: Twenty-three years, three startups, two exits, and the advice still comes down to three unglamorous words. There's no shortcut algorithm for founder resilience, only reps.

Final Thoughts: The Real AI Frontier Isn't the Desk, It's the Field

Ankush's closing perspective: "Anything less than 24 hours a day is part time."

The bottom line: For a decade, "AI for work" has quietly meant AI for people who type. Unfold XR's bet is that the far bigger, far more overlooked opportunity is the 80% of the global workforce that never sits down on the job, from windmill technicians and airport ground crews to geyser repair technicians and construction workers. The company's approach isn't to chase the most impressive piece of hardware, it's to build an instruction and validation layer that works whether someone is wearing a $100 smartphone or a $5,000 industrial headset.

What makes this conversation interesting isn't just the technology, it's the discipline behind the strategy. Ankush deliberately avoided betting the company on any single device category, because in a fragmented, immature hardware market, device lock-in is a founder's fastest path to irrelevance. Instead, Unfold XR chose to be married to the outcome, a human doing their job better and safer, not to the gadget on their face.

With Jio entering the smart glasses category in India and Amazon committing to phase out QR scanners across North America within two years, the runway for this category just got a lot shorter. If Unfold XR's pilots at the airport and beyond keep hitting on time and accuracy, the next couple of years may decide whether frontline AI becomes as ordinary as a QR code scanner, or stays a niche industrial tool for a little while longer.

Q: How can people connect with you and learn more about Unfold XR?

Ankush Jagga: If you're running any frontline business, any field service operations, building in quick commerce, please check out Unfold XR, we have some amazing use cases we're working through.

Final words: The next wave of AI adoption won't be won in a chat window, it'll be won on a rooftop, in an aircraft hangar, on a factory floor, anywhere someone is doing real, physical work with their hands and no time to look something up. Ankush's 23 years of building have taught him that usefulness earns pricing power, not the other way around. For founders building in hardware-adjacent, unglamorous categories, that's the lesson worth carrying forward: solve for the person doing the job first, and let the business model catch up.

---