He fired all his customers. Then built a $1B startup in 2 years. | Jay Madheswaran, Co-Founder of Eve

Most founders obsess over firing employees when their startup hits turbulence. Jay Madheswaran did something far more radical: he fired all his customers. In 2023, with a $3 million ARR AI company serving big law firms, Madheswaran made the decision to walk away from every existing client and rebuild from scratch. Two years later, Eve reached unicorn status by targeting a completely different market with a 10x higher price point.

The decision wasn't born from desperation but from brutal market clarity. Madheswaran realized that the legal AI wave everyone was riding in 2018 had been a mirage. "Back then in 2018, AI does not mean what it used to mean today," he explains. "It was almost like the false start, like the first wave of AI, pre gen AI, was like the false start of AI where you talked about the things you wanted to do and most people couldn't actually do those things, and then all of a sudden now you can."

But even with real AI capabilities finally available, Madheswaran discovered that big law firms were the wrong target. The fundamental economics didn't work. Big law operates on billable hours, creating perverse incentives against efficiency tools. "When we dug into it, it actually seemed like the amount of value AI was going to have for us at the time wasn't going to be that large. We're talking maybe 10% improvements overall for big law," Madheswaran says. "And there's kind of this tension between not replacing billable hours."

The real breakthrough came when Eve shifted from big law to plaintiff attorneys. This wasn't just a customer segment change, it was a complete business model transformation. Plaintiff attorneys don't bill by the hour; they take cases on contingency and get paid when they win. This creates the opposite incentive structure: they desperately want tools that help them handle more cases faster and win bigger settlements.

The market dynamics became crystal clear once Madheswaran understood the core constraint. Plaintiff attorneys aren't limited by case flow, they're limited by capacity. "Something they look at, which is how many cases can an attorney handle at once," he notes. "And that is indirectly tied to how many cases can their marketing engine feed them with, how many cases are they able to kind of move on to the next step so they're not actively working on it, and ultimately, all of that is labor constraint."

This insight unlocked everything. If attorneys could handle more cases simultaneously, they could scale their practices dramatically. Eve built AI that automates the grunt work, document review, case research, settlement calculations, freeing attorneys to focus on client relationships and courtroom strategy. The value proposition was immediately obvious and measurable.

The results speak to the power of finding the right market-product fit. Eve went from a $6,000 annual contract value to $60,000, a 10x jump that transformed unit economics overnight. "ACV was, like, tiny. It was, like, six k or something. And, you know, we were trying to get it to 60 k," Madheswaran explains. The higher price point wasn't just sustainable; it was necessary given the transformational value Eve delivered to plaintiff practices.

The growth trajectory that followed was extraordinary even by Silicon Valley standards. "The first quarter, I think we hit a million in ARR already for that product. And the two months after that, we add another million, and a month after that, we add another million," Madheswaran recalls. The company achieved 800% growth in a single year, reaching 10 million ARR faster than almost any enterprise software company in history.

What made this possible wasn't just product-market fit but what Madheswaran calls "demo shock", the moment when prospects see the AI in action and immediately understand how it will change their practice. Unlike the incremental improvements that characterized the first wave of legal AI, Eve's tools created step-function improvements in attorney productivity.

The lesson for other founders is counterintuitive: sometimes the biggest risk is staying with your current customers. Madheswaran could have optimized his way to modest growth in big law, gradually improving his $3 million ARR business. Instead, he recognized that the real opportunity required abandoning everything and starting over with the right market.

The next time you're stuck optimizing conversion rates or debating feature priorities, ask yourself Madheswaran's question: are you building 10% improvements for customers with weak incentives to buy, or step-function improvements for customers who will pay anything to solve their core constraint? If it's the former, firing your customers might be the smartest thing you ever do.