In My Father's Store, AI Isn't a Chatbot—It's the UPI Scanner

I spend my days at Pulse AI swimming in the alphabet soup of the future. LLMs, AGI, MoE, GANs. I write and edit articles about how autonomous agents are going to restructure global consulting firms, how multi-modal models can design a video game from a single sentence, and how the very nature of work is on the precipice of a seismic shift. Then, I go home to Lucknow.

My father’s kirana store hasn’t changed much since I was a boy. The same tall glass jars of toffees line the counter, their colours a little faded. The air is thick with the competing smells of freshly ground turmeric, Lifebuoy soap, and the sweet dust of Parle-G biscuits. It’s a place of comforting constants. But in the last few years, one small, laminated square has appeared, taped next to the Ganesha idol on the counter. It’s a QR code.

For my father, this is the most significant technological leap of his forty years in business. More than the new shelves he installed, more than the brighter LED lights. As I watch him, a customer buys a small packet of detergent. She points her phone at the QR code, taps a few times, and a small, tinny voice erupts from a speaker box on the shelf: “Paytm par bees rupaye prapt hue.” Twenty rupees received on Paytm. My father nods. The transaction is complete.

In my world, we talk about AI in terms of intelligence, consciousness, and capability. We debate whether a model is ‘reasoning’ or just performing sophisticated pattern matching. In my father’s world, the most advanced intelligence he interacts with daily is this disembodied voice confirming a payment. He doesn’t call it AI. He calls it “scanner.” To him, it’s just a tool, like the weighing scale or the calculator he’s used for decades. He’s not wrong, but he’s also not seeing the whole picture.

And that’s the great disconnect. The AI revolution being sold to us from Silicon Valley stages is one of dazzling chatbots, hyper-realistic image generators, and English-speaking coding assistants. It’s impressive, and it will undoubtedly change white-collar work. But the AI revolution that is genuinely and quietly restructuring the Indian economy is happening right there, on my father’s countertop. It’s unglamorous, it’s invisible, and it’s profoundly more impactful for a much larger number of people.

That simple UPI transaction is the endpoint of a massively complex, AI-powered system. When that customer’s phone scans the code, it initiates a near-instantaneous chain of events across a distributed network. Her bank, my father’s bank, and the UPI switch all talk to each other in milliseconds. But more than that, running underneath this entire exchange are layers of silent intelligence. Fraud detection algorithms, trained on billions of data points, are constantly scanning for anomalies. Is this transaction typical for this user? Is the location plausible? Does the amount fit the user’s spending profile? This isn’t a human in a back office checking a ledger; it’s a machine learning model making a judgment call in the blink of an eye. The sheer scale of UPI—over 14 billion transactions in May 2024 alone—would be impossible to manage without this kind of AI-driven security.

This is not the stuff of breathless tech headlines. A well-performing fraud detection model doesn't give a splashy demo. Its success is measured in the problems that *don’t* happen. For my father, it means he can trust the tinny voice. He doesn’t have to worry about fake currency notes or bounced cheques, two of the perennial anxieties of small merchants from his generation.

Beyond security, every scan is creating a data point. My father doesn’t see a dashboard or analyze trends. He has something much simpler and more useful: a text message at the end of the day summarizing his digital collections. But in aggregate, the data from his store and millions of others like it is a river of information. The payment companies processing these transactions use this anonymized data to understand consumption patterns at a granular level never before possible. They can see which brand of instant noodles is gaining traction in a specific neighborhood in Lucknow or when sales of cold drinks peak during a heatwave. This is market research on a national scale, happening in real time.

This data trail generates another form of practical AI. It creates a financial identity for my father. For decades, he ran a cash business. To a bank, he was largely invisible, his income unverifiable. Getting a small business loan would have been a bureaucratic nightmare of paperwork and uncertainty. Today, he has a verifiable, digital record of his daily sales. Fintech companies, using AI-based credit scoring models that analyze this transaction history, can offer him a small, pre-approved loan with a single tap on his phone. This isn’t a hypothetical benefit; it’s how he financed his last big stock purchase before Diwali. Financial inclusion isn’t an abstract policy goal here; it’s an AI model deciding my father is a good credit risk.

The intelligence flows down the supply chain, too. The distributor from whom my father buys his biscuits and soaps now uses an app on his tablet. This app doesn’t just take orders; it suggests them. Based on past sales data from my father’s store and others in the area, the app’s backend algorithm can predict that he’s likely running low on 10-rupee packs of Britannia milk biscuits. When the salesman visits, he isn’t guessing. He’s working from an AI-generated recommendation. My father just sees a helpful salesman who seems to know what he needs. He doesn't know that the salesman’s intuition is being augmented by a predictive model in a server farm hundreds of miles away.

This is where the Indian AI story diverges so sharply from the Western one. The narrative there is often one of replacement and hyper-optimization. Amazon Go stores eliminate cashiers entirely. Customer service is offshored to chatbots. The goal is often relentless efficiency, which can feel cold and dehumanizing. Here, in my father’s store, the AI is not replacing him. It’s empowering him. It’s taking away the tedious, risk-prone parts of his job so he can focus on the human parts: greeting Mrs. Sharma, asking about her son’s exams, and recommending a new brand of tea he thinks she’ll like.

This AI solved the ‘chutta’ problem—the perennial hunt for loose change that could hold up a queue and fray tempers. It solved the ‘khata’ problem—the informal credit book where he’d jot down what neighbors owed him, a system built on memory and trust that was often a source of minor disputes. Now, if a neighbor is short of cash, they can pay instantly from their phone. His relationships with his customers have become less transactional and more social, which is the very essence of a community kirana store.

When I try to explain this to my friends in tech, I sometimes feel they don’t quite get it. They want to know what model it’s running, what the architecture is. They want to talk about the tech stack. But the truly revolutionary part is not the technology itself, but its application and distribution. The UPI stack is a public digital utility, a platform upon which private companies can build services. This has allowed for a kind of bottom-up innovation that is uniquely Indian. We leapfrogged decades of legacy financial infrastructure—clunky credit card machines, expensive POS systems—and went straight to a solution that works for a tiny stall selling chai as well as it does for a large supermarket.

So when I return to my desk at Pulse AI, surrounded by chatter about the future, my perspective is grounded. The most exciting AI isn’t necessarily the one that can write a poem or create a photorealistic image of an astronaut riding a horse. It’s the invisible, almost boring AI that works silently in the background of our lives. It is the intelligence that secures, verifies, formalizes, and includes.

The next time I visit my father’s store, I know what I’ll see. He’ll be wiping down the counter, arranging packets of chips, and the little speaker box will chirp to life again, announcing another twenty or fifty or a hundred rupees received. That sound, to me, is the true anthem of the AI revolution. It’s not the roar of a super-intelligent machine. It’s the quiet, steady pulse of a billion small transactions, powering an entire economy, one kirana store at a time.