Would you shut down the business you convinced everyone to fund?

New TJC episode is out today!! And guys, today’s guest raised $110 million in less than a month. These days, that might sound like just another Silicon Valley funding announcement. But the business behind this round makes most of its revenue in Latin America. And if you’ve followed the funding environment in the region, you know why I wanted to understand how he pulled that off.

I sat down with Dileep Thazhmon, founder and CEO of Jeeves, about three years after our first conversation in Miami.

And boy, what a roller coaster these three years have been for the company.

In 2023, interest rates were rising, and Jeeves was serving everyone from small businesses to large enterprises. The team realized that some of those customers weren’t generating much revenue—and that other products could actually serve them better.

Dileep made the call to shut down parts of the business he had spent years building.

The company would focus on mid-market and enterprise customers operating across borders. They spent that year rebuilding. He describes 2024 as almost starting from scratch.

I think we sometimes underestimate what that takes. You’ve raised money, hired people and convinced everyone that you’re building something important. And then you have to tell them that parts of what you built shouldn’t continue.

And the outcome was far from certain. Dileep told me there were moments when the team wondered whether they were going to make it.

Today, Jeeves’ revenue is 10X what it was two years ago. The company has fewer employees.

So what happened between those two points?

Dileep describes Jeeves as the story of two startups: the company before AI and stablecoins, and the company after them.

The decision to rebuild put Jeeves in a position to take advantage of both. But neither was as simple as adding a new feature to the existing business.

Take AI.

Jeeves now has four or five people on an underwriting team that previously would have required 25. Its engineering team uses an internal AI system to investigate payment problems and prepare fixes, with a human reviewing changes before they go into production. Dileep says engineering output has increased fivefold.

But what I found especially interesting was how this changed his decisions as a founder.

Previously, he might have made two big product bets in a year and needed one to work. Now, he can make ten and look for two or three successes.

That is a very different way to think about building a company.

You still need to understand your customers. You still need security. But when testing an idea takes a fraction of the time and money, you can try things that would never have made it onto the roadmap before.

The same thing is happening with geographic expansion.

Dileep told me that entering a new country previously cost Jeeves at least $1 million. It meant setting up local card issuing and investing in five or six people.

With its stablecoin-based issuing setup, he estimates that cost is now around $200,000.

In the previous three years, Jeeves had launched in one new country. At the time of our conversation, it was preparing to launch in ten that year.

So naturally, I wanted to know: if all of this is becoming easier for Jeeves, isn’t it also becoming easier for everyone competing with Jeeves?

His answer gets to something much bigger about where fintech is going.

The technology is getting cheaper. The financial infrastructure, licenses, compliance and customer trust still take years to build.

As Dileep put it, you can launch a neobank out of a box. That doesn’t mean you’ve built Nubank.

You haven’t built its brand. Its network. The reasons people trust it with their money.

And this is where those difficult early years at Jeeves start to matter. The company had already spent years building infrastructure across countries. When stablecoins became an opportunity, it had customers, relationships and systems it could build on.

Even then, deciding to launch wasn’t obvious.

The team debated whether enterprise customers actually needed stablecoins. They already had access to dollars. Would they use the product? What problem would it solve for them?

Dileep eventually made the call: they had four weeks to build it before he would announce it at a conference.

He called it his hardest decision of the previous twelve months.

And yet, when I asked what drove adoption, his answer had very little to do with customers getting excited about crypto.

Jeeves serves companies including Burger King, BMW, Lululemon and H&M. These are finance teams trying to pay suppliers, manage spending and keep a business running across countries.

They want to know: will the money arrive? Is the rate good? Can I trust you?

Dileep’s version was even more direct: “Am I not gonna get fired because, you know, I used stablecoin?”

Honestly, that probably explains enterprise adoption better than a lot of crypto presentations 😂

For Jeeves customers, the steps involved in sending a payment stayed the same after stablecoins were introduced. The company changed the infrastructure underneath a product they already used.

Through all of these changes, Dileep says the ambition stayed the same: build a global business bank.

Most of Jeeves’ revenue comes from Latin America. But the companies it serves have suppliers, customers and offices elsewhere. Brazil, Mexico and Colombia are a starting point for that ambition.

Which brings us back to the $110 million.

By this round, Dileep was pitching a very different business from the one we had discussed three years earlier. He also wanted investors with experience in the direction Jeeves was now taking.

CoinFund led the round, joined by investors including Coinbase Ventures, ParaFi and Wintermute.

But even with that growth, and even with the round coming together so quickly, there were plenty of nos.

I told him I understand why founders struggle with fundraising. Asking people for money is hard. And hearing no can feel so personal.

His response was immediate: “You can’t take it personally.”

The investor’s job is to say no most of the time. Your job is to keep going and find the right fit.

He was equally direct about how to run the process: don’t give away too much information before you’re ready, bring investors in on a structured timeline, and treat fundraising like the full-time job it is.

And do your homework on the actual person writing the check. You could be working with that partner for ten years—through the kind of decisions Jeeves had just spent three years making.

A recognizable fund name tells you very little about how that relationship will feel when things get difficult.

So toward the end, I asked Dileep a question that gets much less attention than how much someone raised or what their company is worth:

What is the most expensive term in a term sheet that isn’t valuation?

He didn’t hesitate.

He named something founders can agree to in one round that makes every round after it harder—and explained why he would give up other terms to avoid it.

What was it?

1. The experience that made you a great executive five years ago can work against you today.

Dileep questions how much yesterday’s scaling experience is worth when AI changes the resources needed to build a business. Executives arrive knowing how many people a team needs, how long a project takes and which tasks require outside software. But what happens when those assumptions are outdated? That’s part of why Jeeves invests in interns. One built an internal HR system the company had been using for 18 months, including for performance reviews. As Dileep puts it, interns “don’t have that baggage.” The hiring challenge is finding experienced people willing to reconsider the playbook that made them successful.

2. Managers who only manage may have a much harder time justifying their role.

Dileep expects leaders to become what he calls “super ICs”: individual contributors who manage people but can also do the work themselves. That expectation extends to the C-suite. He wants leaders testing new tools, solving problems and staying close enough to the work to step in when something isn’t getting done. What’s interesting is how this changes the meaning of seniority. Moving up the organization doesn’t mean moving further away from execution. As AI lets individuals accomplish more, a leader’s own ability to build becomes more valuable.

3. Jeeves’ breakthrough in Brazil came from solving a very specific problem in corporate travel.

Dileep says Jeeves grew almost 5X in Brazil over the previous year, with corporate travel playing a major role. The company built software for travel providers whose customers needed large numbers of single-use virtual cards. Those cards help reduce fraud and make reconciliation easier; one company might require 100,000 or 200,000 cards a month. I love how concrete this is. The opportunity was in understanding a local industry’s operational problems well enough to build something customers actively pulled into their businesses.

4. You don’t have to replace a bank to build a huge business around its customers.

Dileep sees Jeeves as an operating layer that can sit above existing banks and make life easier for finance teams. His example: a company with offices across several countries might only see its consolidated spending weeks after the month closes. Bringing that information into one place, in real time, is already a meaningful improvement. He says 80% of Jeeves customers use more than one product line, allowing the company to grow within relationships it already has. His broader point is that financial services is large enough for many substantial businesses serving different needs.

5. A CEO’s job is to make decisions with 50% of the information—and know which missing facts actually matter.

Dileep says much of his day involves decisions that have already worked their way through the organization before reaching him. By then, there are competing opinions, incomplete information and a choice someone still has to make. His approach is to ask whether he has enough information to decide. If one specific data point could change the answer, he looks for it. If he already has what he needs, waiting simply postpones the decision. That distinction is useful: what exactly do you still need to know, and would knowing it change your call?

Olga xxx

🎙 The J Curve  is where LATAM's boldest founders & investors come to talk real strategy, opportunity and leadership.