SabPaisa Becomes India’s First AI-First Fintech Company: A Conversation with Abhimanyu Jha

An Exclusive Interview with Mr. Abhimanyu Jha, Co-founder of SabPaisa, an “AI-native” company

Abhimanyu Jha, Co-founder of SabPaisa, leads the charge in transforming fintech into AI-tech. As an “AI-native” company, SabPaisa integrates intelligence and innovation to revolutionize digital payments.

In this interview, Jha discusses how AI is driving smarter, secure, and inclusive financial solutions for a rapidly evolving digital economy.

How did the journey of SabPaisa evolve into what is now called an “AI-native” company, and what does that term truly mean in practice?

Abhimanyu Jha: The evolution happened in stages; and it started from an unexpected place. Last year I was writing a book.

The book was on an almost hundred years old, one of the most famous, yet still mostly answered questions about the role of mathematics in our universe, a question that was also asked by Einstein himself: why is mathematics so effective in explaining our universe.

The question became famous when another of 20th century’s top scientists, the Nobel Prize winning physicist Eugene Wigner, wrote a widely popular essay on the question in 1960 titled “The Unreasonable Effectiveness of Mathematics in the Natural Sciences”. 

While doing my research on the book, I went through perhaps more than a thousand research papers across mathematics, physics, biology, philosophy of science, philosophy of mathematics etc; and here AI tools like chatGPT and Claude helped me in ways I hadn’t anticipated.

They took me to places I’d never thought to look, made connections I wouldn’t have seen, as long as I knew how to guide them properly.

I was astonished at the power you could unlock in these tools if only you knew how to talk to them properly and nudge them back on the desired paths when inevitably they hallucinated away at some point.

That experience became the foundation for what we now call Moneyball AI at SabPaisa (the name is borrowed from the story of a Hollywood movie starring Brad Pitt called Moneyball). 

In April this year, we made it mandatory for every team member to use AI to solve their problems. Around the same time, Shopify’s founder Tobi Lutke published a memo that you can title “Reflexive AI,” the idea that people should turn to AI reflexively, like they previously turned to Google. 

But we went much further than what Tobi had suggested. Just a couple of weeks after Tobi’s memo, I wrote the Moneyball AI memo to all our team members saying that Reflexive AI wasn’t enough, that turning to AI reflexively was still treating AI as an assistant, an intern.

Now top AI models had become so powerful thanks to Reinforcement Learning and Reasoning advances – one of the first examples being the Deepseek model that went viral in January – that people who really knew how to talk to AI well and knew when to trust it and when to bring it back on the desired path when it hallucinated, those people could go beyond treating AI as an assistant and start using it as a guide who could take you places you had never been, as if it was your teacher, your guru.

Here’s a simple to understand comparison between Reflexive and Moneyball AI drawn from real life: Imagine trying to drive across Delhi during peak rush hour.

Let’s say you need to go from India Gate to Lal Quila. You could invest enormous mental energy analyzing traffic patterns, weighing different routes, guessing about unpredictable delays. It’s a really complex problem, one you probably won’t solve perfectly on your own. 

Or, you delegate. You simply tell Google Maps your destination and trust its algorithms, which are constantly fed with vast amounts of real-time data, to find the optimal route and drive accordingly.

You effectively offload the how. This simple act frees your mind to think about why you’re making the trip, what you’ll do when you get there, or even just to observe the city around you. 

And then you climb a step higher as the AI system becomes better and self-driving mode arrives. You are not even driving anymore. You have just told the AI you need to go from India Gate to Lal Quila and it will do both the route calculation and the driving while you sit back and relax.

You can even tell it whether you are in efficiency mode and have to travel as quickly as possible because you have to attend a meeting or whether you are in shopping or sightseeing mode and would love to take a tailored even if slower route more adapted to your specific needs e.g. take a route on which there are best sweet shops because you want to eat awesome Rasgulla today, or take a route that has the best blazer shops because you want to buy a smart blazer for your partner or boy/girlfriend.

Essentially, you just tell it the objective, the initial point, and the final destination, and the intelligent driving system finds/plans out the best route for you, and in the agentic future, might even execute it for you.

Driving a car rather than walking to get from India Gate to Lal Quila is like using Reflexive AI to get the job done at a super-fast speed; but still you are in total control and the AI is just a super-fast machine, an instrument, to get the job done.

Moneyball AI is just telling the car I want to go from India Gate to Lal Quila through a road that has the best Rasgulla shops and letting it do everything else such that AI is not just a superfast intern anymore, it’s practically your teacher showing you a new route that you never knew before.

Reflexive AI is thus more of a simpler tool, instrument change; Moneyball AI is a fundamental process, behavior change, and is thus both much more difficult and much more rewarding.

What about the fear that letting this happen will make us lazier, mentally weaker or make us lose our skills? 

The counterpoint to that is do you really gain much mental strength or driving skill in driving across Delhi during rush hour?

If you really want to build your cognitive capacity and driving skill, won’t it be a better idea to take part in driving competitions or challenges like Desert Storm or Dakshin Dare, or if you really have the guts, the Rainforest Challenge?

Essentially, go ahead and attempt stuff that is either difficult for AI, or even better, attempt stuff that has never ever been attempted before? Let AI decide on and do at least the routine, everyday stuff so that you have time and energy to attempt extraordinary things.

Delegating complex analysis, optimization problems, or routine tasks to AI can therefore liberate our human minds. It allows us to focus our energy on the truly novel challenges, the creative breakthroughs, the high-level strategic thinking, essentially exploring the territories not yet mapped. 

What makes an organization truly AI-native? 

Think about the last time when a similar technological revolution swept the world off its feet. It was the Industrial Revolution that spanned from the mid 1700s to the early 1900s; and which came about with the invention of the heat engine by James Watt. 

When cars were first invented by Karl Benz in 1885 towards the end of this revolution – even then, horses, bullocks, donkeys, mules or camels were still the dominant means of transport. But when Benz invented the car, he didn’t create an iron horse with an engine inside.

He completely reimagined the situation by putting the heat engine, the driving force of the entire industrial revolution (along with electricity and telegraph), at the center of the design.

That’s what we’re doing at SabPaisa. We have understood we are at the beginning of another technological revolution, this time driven by AI. And so we’re putting AI at the center of our design and redesigning everything else around it.

Everyone is becoming a developer to some extent. All our specialized developers are becoming full-stack. Product and engineering have merged into a single team. These are fundamental reimaginings of how an organization might operate in the AI age.

How is the transformation of the entire workforce into AI developers being implemented, and what is one real-world example that shows how it works?

Abhimanyu Jha: The most dramatic example might be my own. In June, I stood before the entire company and took on a challenge: I would build a fully functional application in three days, despite never having coded professionally in my life. The last time I touched anything resembling code was learning Visual Basic in a classroom more than twenty years ago.

Between June 13th and June 16th, I developed a desktop application called AI Workflow Analyzer that has around 1,300 lines of code (minus tests and dependencies) with proper version control and integrated tests.

Every single line down to the last letter was written with the help of Claude and Gemini – all I did was give commands, copy/paste where required, and take screenshots for debugging. 

On the third day, right at the deadline, I demoed the application to the entire company. One hundred percent of the code was AI-generated, but I was the architect, the guide, the one steering it toward the goal.

What happened next was even more interesting. I challenged the non-developers in the company to emulate me and multiple non-developers took up the challenge over the following month and successfully created their own applications.

We’ll soon have a felicitation and prize distribution ceremony for them. This changed how everyone thinks about their capabilities.

The broader implementation is systematic. We’ve mandated that everyone must engage with development to some degree, because the walls between functions are dissolving.

When a marketer understands how software actually gets built, when a finance person can prototype a tool they need without waiting for engineering, when everyone can speak the language of creation, the entire organization moves differently.

Roles shift like an actor shifts between characters. You might be working on marketing in the morning, contributing to product development during the day, and analyzing financial data in the evening.

This transformation is real and happening now at SabPaisa. We’re living it and documenting what we learn so others can follow.

What inspired a non-coder to start coding, and what lessons did that experience offer about leadership in the AI era?

Abhimanyu Jha: For the past six years at SabPaisa, I couldn’t take on certain roles because I lacked software development skills. That constraint shaped what was possible.

But after seeing what AI could do while researching my book, how it could guide me through complex scientific literature when I knew how to work with it properly, I realized something had fundamentally shifted.

The question became: if AI has become this powerful, what becomes possible that wasn’t before? The only way to truly know was to test it myself. So I made it a public challenge, with a deadline, in front of the entire company.

When the founder stands up and says “I’m going to do something I’ve never done before with these new tools,” it gives everyone else permission to reimagine their own boundaries.

The lessons are significant. First, the constraints that defined our capabilities for decades are evaporating. Second, leaders must embody transformation, showing the way by doing. Third, and perhaps most important, this changes how we think about education and career paths entirely.

If roles can shift like characters in a play, becoming fluid identities, what does that mean for how we prepare people? What does it mean for MBA programs if career paths become dynamic across domains? At SabPaisa, we’re already operating in this new reality.

We believe what we’re doing should eventually be studied as case studies at places like Harvard or Wharton, because we’re prototyping answers to questions the education system hasn’t even fully asked yet.

The coding challenge was about proving that the fundamental nature of capability itself is changing, and founders/leaders need to recognize and demonstrate that shift.

What connects the ideas behind Moneyball AI and the books inspired by Einstein’s questions, and why combine philosophy with technology in those writings?

The connection is profound and it’s about understanding what intelligence fundamentally might be, whether it’s human intelligence or artificial intelligence.

As I mentioned earlier, Eugene Wigner (and Einstein) asked a question that’s haunted scientists and philosophers for decades: why is mathematics so unreasonably effective at explaining our universe? Why does math, which we create in our minds, work so perfectly to describe reality? 

I believe, and perhaps I am the first person in the world to do so, to discover that perhaps there is a link between Wigner’s question and the ontology of AI.

That if you want to truly understand what AI is and what it’s capable of, and more importantly what it can and can’t do, you first have to explore Wigner’s question, which perhaps tells you about the fundamental role of mathematics in human life and universe, and further helps you understand similar questions like “what is intelligence?” – whether carbon-based or silicon-based, biological or artificial.

That’s a quadrillion-dollar question being asked by many of the top minds in the world. 

Right now, people are both wildly optimistic and deeply fearful about AI, often without really understanding the nature of the intelligence they’re dealing with.

Some think it will destroy the world like in Terminator movies. Others think it will solve every problem. The truth is we can’t control what we don’t understand. And we can’t avoid dangers that we can’t comprehend.

One of my important goals for the coming year (and SabPaisa will also work towards it so we can all do something important for the world as a company) is to draw the world’s attention, including top AI minds and Nobel Prize winners, people like Fei Fei Li, Demis Hassabis, Geoffrey Hinton, Sam Altman, Ilya Sutskever, Dario Amodei and others, to this question about unreasonable effectiveness of mathematics.

Forget the answer I’ve found, whether it’s right, wrong, mostly right and a little wrong, partially right and partially wrong… whatever. I want everyone to think about this question, because in my opinion wrestling with it is essential for understanding AI.

Moneyball AI, our approach to using AI as a guide, emerged directly from this deeper thinking. It’s grounded in understanding that AI represents something closer to a force of nature than a simple tool. And like any force of nature, like a river that can both create floods and give life, it’s a double-edged sword.

The philosophical work informs the practical implementation, and the practical implementation reveals which philosophical questions matter most.

That’s why we’re connecting our company’s identity to these deeper questions. Every major technology shift, the Industrial Revolution, electricity, the internet, forced us to ask fundamental questions about human capability and purpose. AI is no different.

At SabPaisa, we’re contributing to humanity’s understanding of how to work and think alongside genuinely transformative technology.

What makes India uniquely positioned to lead the world in AI, and what changes are needed for that vision to become reality?

Abhimanyu Jha: I want to be clear. I’m not claiming India is currently leading the world in AI in the way Silicon Valley (in software) or China (in hardware) is. But India has distinctive advantages that could position us strongly for the AI 1st future if we make the right moves.

First, consider the resource constraint paradox. I will take SabPaisa’s example here. We’re a bootstrapped company competing against larger players who’ve raised tens or even hundreds of millions of dollars. Until now, we were still small because of resource constraints. That forced us to think differently from day one.

We couldn’t win by outspending competitors on traditional marketing/sales or by hiring the largest development teams. We had to find fundamentally different approaches, which is exactly what led us to reimagine our entire organizational structure around AI.

This scarcity-driven innovation is something India understands deeply. We’ve built world-class technology companies with a fraction of the resources available in Silicon Valley.

When AI democratizes capability, when a small team with the right approach can suddenly accomplish what previously required massive resources, that plays to India’s strengths.

Second, we have massive scale and diversity. When you’re building AI systems that will serve billions of people across different languages, contexts, and use cases, having that complexity in your home market is an advantage.

The payment infrastructure we’re building has to work for government departments in Karnataka, universities across the country, and businesses of every size. That complexity forces sophistication.

Third, and perhaps most important, is what I’d call intellectual audacity combined with practical grounding. Indian technologists aren’t afraid to engage with big ideas, to ask fundamental questions, while simultaneously being deeply practical about implementation. That combination is exactly what’s needed in the AI age.

But for this potential to become reality, we need changes in mindset. We need to reimagine from first principles.

We need to be willing to make bold bets on organizational transformation. And we need to document and share what we learn transparently, so India becomes known as a place where the future of AI-enabled organizations is being actively prototyped and proven.

What are the biggest challenges being faced in building or adopting AI today, both at an organizational and societal level?

The biggest challenge is also the most fundamental: people are treating AI like it’s optional, or like it’s something they can adopt slowly when they feel ready. That’s a dangerous misunderstanding.

AI is a force of nature. I don’t mean that metaphorically; I mean it quite literally. Yes, AI was created by humans, but the principles it runs on, the mathematics underlying it, those are laws that govern our reality.

When humans invented trains, we created the machine, but the thermodynamic principles that make engines work are natural laws we discovered. 

AI is similar. Forces of nature don’t wait for you to be comfortable with them. They don’t stop because you’re ready.

It’s not in your hands whether this force of nature will affect you or not. At best, it’s in PM Modi’s hands or Mr. Trump’s hands or Mr. Jinping’s hands. And likely it’s not even in their hands. PM Modi won’t stop it because Mr. Trump or Mr. Jinping may not.

And then what happens to India? Same logic will be running through the minds of Mr. Trump or Starmer or Xinping or Macron too. Nobody’s going to stop this. Nobody perhaps can stop this. At least I don’t see that happening on the horizon, not anytime soon. 

So the only question or challenge is whether we’ll transform intelligently or chaotically, whether we will do it voluntarily or will it be thrust upon us unwillingly. The latter is a much worse option.

At the organizational level, the biggest mistake I see is what I call the “steel horse syndrome.” When cars were first invented, Benz didn’t try to make steel horses with engines inside them. But that’s exactly what most organizations are doing with AI right now.

They’re taking their existing processes, their existing workflows, their existing organizational structures, and trying to squeeze AI into them. The question “Where can we add AI to this existing thing?” misses the fundamental point entirely.

At SabPaisa, we are restructuring everything. Everyone is becoming developers to some extent, specialized roles becoming fluid, teams merging in new ways.

That’s uncomfortable. Change always is. But the alternative is falling behind competitors who do reimagine from first principles.

At the societal level, the challenge is more complex. AI is a double-edged sword, like a river that can cause floods but also sustain life.

We can’t pretend the dangers aren’t real. I’ve talked about potential problems arising from AI myself in a close to 15000 word essay that I wrote two years ago in July 2023 including its effect not only on jobs but even on relationships and marriages.

I took the essay down because of backlash on Twitter/X because I didn’t want to be seen as a Luddite being the founder of a technology startup.

Anyway, we also can’t let fear paralyze us into inaction. The world needs people who are actually implementing AI at scale, learning what works and what doesn’t, and sharing those learnings transparently so everyone can benefit.

One specific challenge troubles me deeply: education systems are still preparing people for a world of fixed roles. You study finance, you become a finance person for life. You study marketing, you’re a marketer forever.

But at SabPaisa, we’re already seeing roles shift like an actor shifts between characters in different plays. With AI, you might be doing marketing work in the morning and contributing to product development in the afternoon.

If educational institutions don’t adapt to prepare people for this fluid reality, we’ll have a serious mismatch between how people are trained and how work actually happens.

The overarching challenge is speed. The gap between people and organizations who are adapting quickly and those who are moving slowly is widening every month. At SabPaisa, we recently delivered a solution for a Karnataka government department in 24 hours.

That speed advantage compounds. If you fall behind in speed, you may never catch up. That’s why I’m so emphatic about this: ignoring AI or moving slowly on adoption is existential.

What is the single most important mindset shift entrepreneurs need to make while building in the AI era?

Abhimanyu Jha: The fundamental shift is this: AI needs to be your foundation, the center around which you build everything else.

When we started implementing AI at SabPaisa, the initial instinct was everyone in leadership roles asking “Where can we use AI to make our existing processes better?” That’s thinking of AI as an enhancement, an optimization tool. It’s useful, certainly, but transformation requires going much further.

The real shift happens when you ask different questions. If we were starting this company today, with current AI capabilities available from day one, what would we build? How would we organize? What would be possible? Those questions force you to reimagine.

This mindset shift has a cascading effect. When you redesign workflows around AI’s capabilities, everything changes. Roles change. Team structures change. Even your fundamental value proposition might change.

What that looks like in practice at SabPaisa: we mandated that everyone engage with development, because the distinction between who builds and who uses is blurring.

We merged product and engineering teams, because the separation made less sense when AI enables rapid prototyping. We told everyone they needed to become comfortable treating AI as a guide, someone who can take you places you haven’t been before, if you know how to steer the conversation properly.

The second part of this mindset shift is about speed and inevitability. Entrepreneurs need to internalize that AI transformation is happening right now. Every month you delay reconceiving your organization around AI, your competitors who have made that shift are pulling further ahead.

The speed at which things move has fundamentally changed. We delivered complex solutions in 24 hours that previously would have taken much longer.

That kind of acceleration changes what’s possible in terms of customer responsiveness, iteration speed, and competitive positioning.

The third element: transparency and sharing. We’re documenting the journey and sharing what we learn. Why? Because this creates differentiation that competitors can’t easily copy.

They can match our marketing spend. They can try to copy our products. But they can’t manufacture the thought leadership and industry recognition that comes from genuinely pioneering new approaches and sharing the learnings openly.

If I had to distill this into one sentence: Treat AI as the foundation you design your entire organization around; and do it immediately, because the transformation is already underway.

From fintech to AI-tech, Abhimanyu Jha’s vision for SabPaisa marks a new era of intelligent financial transformation.

His approach showcases how AI enhances trust, accessibility, and innovation, positioning SabPaisa as a leader shaping the future of payments with purpose and technology.

Read more Success stories of Indian entrepreneurs, Women Entrepreneurs & startups stories at SugerMint. Follow us on Twitter, Instagram, Facebook, LinkedIn