Lance Glinn:

Welcome into another episode of the Inside the ICE House podcast. Today's guest is Dileep Thazhmon. He is the founder and CEO at Jeeves. Dileep, thanks so much for joining us Inside the ICE House. Happy to have you here.

Dileep Thazhmon:

Thanks, Lance, for having me. I'm excited to be here.

Lance Glinn:

So let's start at the beginning for just a couple minutes. The days before Jeeves existed, what was the moment or the set of experiences that made you realize that there was a real gap in financial infrastructure for business? And how did that initial insight help you in saying to yourself, "Hey, this is a company I want to start. This is a problem I want to solve."?

Dileep Thazhmon:

Yeah, so actually, my last company before Jeeves was based in New York and we had an office outside the US as well. And a lot of the problems that I saw in that company is what we're trying to solve with Jeeves. So what that means is, if you are a company that operates in more than one country, US plus Mexico, US plus Brazil, your financial stack is pretty convoluted. You need one provider for a corporate card in country A, you need another provider for a corporate card in country B, a different provider for payments in country C, and so on. And so to me, it was an actual lived problem, because in my last company, it would take us about 30 to 45 days after the month closed to get an idea of where our cash was. And this is a common problem for most companies that operate internationally, most companies that operate in more than one or two countries.

And then that idea was accelerated, really, by COVID, because after COVID, companies are now global by default. If you have one contractor in another country, you have one employee in another country, you have a vendor you need to pay in a third country, by definition, you're a global company. And so that's been a big accelerant for the growth that we've seen at Jeeves. But the problem, to me, was personal, it was lived. It was actually based out of New York that I saw that for the first time. And that's driven a lot of what Jeeves has become today.

Lance Glinn:

And Jeeves has scaled and had really great growth. How do you stay true to your roots, that original vision that you had, while that scale happens? Because I think it's very easy for leaders, for founders, for CEOs, whatever total you want to put onto it, it's very easy to, as you grow, move away from what got you there. How do you stay rooted and make sure that that problem that you were originally trying to solve remains in your mind and in your focus?

Dileep Thazhmon:

So two thoughts there. One, I think it's really important for founders and CEOs to be on the front lines. And you can ask anybody in my team, my company, I pride myself on being the fastest to respond to customers. Because they don't give an F about the company, about me, they'll tell you exactly what the issue is. And if you hear it from them, it's unfiltered and it really helps you keep the focus on what matters, which is, really, the customers. They're really my boss at the end of the day. And the more you can interact with them directly one-on-one, the more you keep that touch onto what is important for them. And if it's important for them, then it's important for you. And if it's important for you, then it's important for them. So that, I think, one, is just having that touchpoint with the customers.

And then two, what's really interesting now is a lot of the technology itself is changing. So the problem hasn't changed. The problem is still, finances are fragmented, banking is country and currency specific, companies are global by default. How do you consolidate that? But the technologies to get there has changed. So to your question of how do you stay grounded? The question is, what are you trying to solve? And so for me, that solving hasn't changed, which is our goal, is to build a global business bank and to make it so that if you're a company that operates in more than one country, you only need to have one partner. That's the goal. And the technologies, the stacks, the stablecoins, AI, it just helps you get to that point faster.

Lance Glinn:

Is it hard though with that ever-changing technology to, one, change alongside that ever-changing technology? But also, to make sure that you stay a step ahead of the customer? Because I definitely think you never want to obviously fall behind. That's obvious. You don't want to fall behind with your customer needs. You, at the very least, want to stay level with where they are and make sure that as they adjust, you adjust. But I think the great companies are able to be that one step ahead, or maybe even two or three steps ahead, so that they know what the customer will want before the customer wants it.

Dileep Thazhmon:

Yeah, so it's a really good question and something that is, I would say, pretty nuanced to figure out what that balance is. So I'll give you an example. Obviously, everybody now knows about AI and AI agents. But the reality is, outside the US, where... We operate in 25 countries, our biggest markets are Brazil, Mexico, Columbia. We just launched in Argentina. And the reality is, in some of those countries, there's still not enough trust in using agents to do purchases. If you're the CFO of the company, do you really trust an agent to go buy a ticket for you or is it going to F it up? And so, to me, in theory, that idea is amazing, and we support it. We have full CLIs, MCPs. You can do it agentically, you don't need to be logged in with a human loop. Actual usage for that requires building trust into the system.

And I think there's a good example, where the technology is ahead of actual usage. And to get people to use that, you actually need to teach them and evangelize what it is that you're actually building. And a lot of times companies build it, and it's like, "Oh, if we build it, they will come, because we figured out three steps ahead of where they are." But the missing step is, you actually need to train them. You actually need to teach them, you need to evangelize, why is it good for you, as the CFO, to invest a week in learning this new product sequence, because then it's going to save you three weeks the next time. So that's the gap that I see, where we are investing a lot more in the teaching and training side, independent of the speed of the technology development. Because if the technology is fast, but you're not teaching people how to get to them, you're still going to have a gap because no one's going to know how to use your new technology.

Lance Glinn:

And besides the education point, I would think what also is necessary is not just to educate people on how to use it, but to allow them to trust it. How are you going about doing that? Because I would say that's probably even harder than just the education. Where it's one thing to say, "Here's how you do it," but you could tell me how to do a million different things, that doesn't mean I'm going to trust those million different things that you told me and taught me how to do.

Dileep Thazhmon:

Yeah. And this is very important to what we do. So our core platform targets enterprises. So it targets companies, like BMW, Lululemon, H&M, that need to use us internationally in multiple markets. And so as you are calling out, actually the number one thing is trust. Because if they don't trust you, it doesn't matter if you have stablecoin or AI, you're not going to get the chance to win their usage. And so the way we handle it, we say, "Hey, just test us out. Try it out, see how it is." And, ideally, we have enough of your attention span that it's better than what you're using today, that you want to bring more spend over, you want to bring more cards over, et cetera.

And so, one thing that we track, from a KPI perspective, is what we call revenue retention, which has increased about 385% over 12 months. So that means if a company starts on day one, by day 365, they're spending 385% more than they were on day one. And so that, to me, is a good indicator that the product itself is resonating and they're bringing more and more of their workflow of their sequences onto Jeeves. And so these KPIs help guide, for me, as a CEO, whether it's actually going in the right direction or if it's just noise.

Lance Glinn:

Yeah, absolutely. So the world, through all this ever-changing technology, is more connected than really it ever has before. Yet, as we talked about earlier in our conversation, and one of the impetuses for why Jeeves is where it is right now is because of the still fragmented financial infrastructure system. Why do you think that across countries, currencies and banking systems are still so separate? And even if areas, like communication, like software, so on and so forth, are becoming more connected, why do you think financial infrastructure hasn't gotten there yet?

Dileep Thazhmon:

Because I don't think the incentives have aligned for financial infrastructure to get to a point where it works across countries. So take the US for instance. US is a large single market with one currency, massive economy. Candidly, if you're starting a company and you win in the US, that's a multi-billion dollar company. So you don't have the incentive to be like, "Okay, let me go figure out how to make this work in Mexico." But that's not the case internationally. If you are in Mexico and you're growing, you have to go to Columbia. You have to go to the US. And so the incentives then become, if you're a company that operates again in more than one country, how do you get to a point where you have an infrastructure that works in more than one country? And the big realization that I had for Jeeves is for us to win in this market, we operate in 25 markets, we have to own the infrastructure stack directly.

And so we have our own stack that plugs into these local banks, plugs into local payment rails, abstracts all of that information, and sends that back to you as a user. So for you, as a user, as a CFO at an international company, you just have to log into Jeeves. We can give you cards in 25 countries. We can do payouts to 190 countries. We have multiple partners, but all of that is abstracted out for you. So you get almost one login that works across the board. If that makes sense?

Lance Glinn:

Absolutely. And so it's interesting because it seems like Jeeves is solving a problem that's pretty obvious.

Dileep Thazhmon:

Obvious, but difficult.

Lance Glinn:

Obvious but difficult, for sure. But I guess my question is, if the problem is pretty cut and dry, and I'm not saying that in, "It's an easy solve," way, but if the problem is, let's say, obvious, why has it taken, in your mind, Jeeves to fill it? Why hasn't anyone seen it 10, 15 years ago and tried to solve it?

Dileep Thazhmon:

So if you want the honest answer, it's probably because I don't have a banking background or don't have a FinTech background. So I think if I did, I would also be very daunted by the problem. And the way we're solving it is very different from how a lot of companies look at it. So usually if you're starting a FinTech, what you do is, you go very deep in one country, like the US. You build there for three, four, five years, and then after that, you're like, "Okay, let me go to the adjacent country and build there deep as well." What we have done that's different is, from day one, we have built it to be very modular. And so what that means is, if you look at card issuing, which is our core product, which is, we can give you corporate cards in 25 countries, what we've built scales across 50 countries.

So today we only operate in 25, but tomorrow, if I wanted to launch in Indonesia, I could spin up Indonesia in about three to four months. And that's the difference where we've built a lot of the stack. And I can go into a lot of detail, but the core idea is that the stack itself is very modular. It's not super deep in one country, it's more a layer that sits across multiple countries, and that gives us the ability to launch faster. And again, it's really mainly because I didn't have a FinTech or banking background that I didn't know some of the nuances that come with understanding the full banking stack. And now we've made that jump, where we've made the model work. And so now for someone else to do it, they have to go invest in the same things that take three to four or five years to get licenses regulated, card issuing across countries. Those things take time.

Lance Glinn:

And so I want to use Indonesia as an example, since you just brought that country up. Obviously, like you said, you have the stack to be able to expand quickly if you wanted to pass those 25 countries that you're already in. What is the process? What is your own personal philosophy, process, your leadership philosophy, when it comes to when the right time is to expand, what the right place is to expand to? How do you determine, "This is where we want to go. Here's when we're going to do it. Here's how we're going to do it. Here's where we're going to do it."?

Dileep Thazhmon:

So our goal is to build a global business bank, so like a Revolut for enterprises. If that makes sense? So really focusing on enterprises covering 25, 35, 45, 50 by 65 countries. So my goal is to cover as much of the globe as possible. Now that said, it's a really good question of how do you decide which country? And so we have three criteria that we look at. So one, it has to be a country that has FinTech V1 infrastructure. So what does that mean? It has some version of open banking. It has some version of invoice tracking. Because that helps us make a decision on whether this company that's coming to us, this enterprise, is a good company to work with for providing corporate cards, et cetera. So that's one, which is just we want financial infrastructure.

The second one, is this a country that has a central banking regime that is open to FinTech? So some countries that's hard. We would never operate in China. We just don't have the knowledge. It's not a market we're familiar with. But if you look at Mexico, you look at Brazil, the central bank is actually very, very forward-thinking. And so that's the second one. And then the third one is just follow the money. Where's the money going? So if you look at Mexico, it's a huge partner of the US for trade. There's so much payment volume back and forth. It made sense for us to start in Mexico. And then now Brazil is growing incredibly fast and it's going to be our largest market next quarter. Same thing. It has these three criteria. We're like, "This is a jackpot. Let's go deep into this country."

Lance Glinn:

So I want to pivot the conversation a little bit onto the impact of stablecoins. And I know you've been looking and focused on stablecoins as part of the infrastructure. How do you explain to current or prospective customers what role stablecoins can play in financial operations?

Dileep Thazhmon:

Yeah, so it's interesting because the way we position stablecoin, we actually don't sell stablecoin itself, we sell the outcome. So what does that mean? Let's say you're a company in Mexico and you need to do a payment to Europe. Today, if you were doing that payment, you would log into the app, you would put that information, you would get the money that you have, send it to one bank, send it to a second bank, pay it out. It'll probably take you a day, maybe a little bit more than a day. Now, with stablecoin, it still looks exactly the same, meaning you still come in, you still do the payout, you put your information for SEPA to pay out in Europe. But what we do is, we collect the funds in Mexico. We then settle on USDC as the ledger. And that's what's much, much quicker than having one to two correspondent banks in between. And then we pay out on the other side.

So it's a classic stablecoin sandwich, if that makes sense, where we have the pay-in that's local currency. In this case, Mexican Peso. We have the payout that's local currency. In this case, that's euros. But the ledger, the inside is now on USDC or USDT, whatever it is that you want to use. And so that is what gives it so much more speed. But the outcome is still that your payment is there faster, your payment is there cheaper. And when we sell to CFOs, you have to sell the outcome. So long way to answer the question is, we don't really sell stablecoin, we sell the outcome, which is faster, cheaper, but you can trust us, and it just so happens to run on stablecoin. If that makes sense?

Lance Glinn:

Yeah, absolutely. And is there anywhere specifically you're seeing the most compelling real world use cases with that technology? Is it supplier payments? Is it treasury? Is it settlement? Where are you seeing the location?

Dileep Thazhmon:

Yeah, so it's a good question. And so if you think about the US, where I think it takes the most, what's the word, explanation for what a stablecoin is. And if you think about why, it's because stablecoins are usually pegged to the US dollar, and you already have access to the US dollar. But for a lot of people, especially outside the US, they don't have access to the US dollar, and you have currency volatility. If you're in Argentina, you're in Venezuela, your currency is going up and down. You get paid today, tomorrow that's worth half of your money. And so that is a huge motivator for using stablecoin. So in Argentina, 60% of the population use stablecoin. It's one of the largest stablecoin adoptions in the world. And the reason is, it's a learned experience. Today, if I have 10 Argentinian pesos, tomorrow that's going to be worth five. So I'd rather put it into dollars.

So big driver is exactly that, which is protection against currency volatility. And then the second one that we are seeing is that it unlocks these countries that we are trying to launch into. So as an example, we're launching this Argentinian stablecoin product, where you can pay an Argentinian peso. As soon as it comes in, it converts to USDC. So now you get protection against the currency volatility. And on the USDC, we can issue the corporate cards that these enterprises need. So what that means is, now instead of 25 countries, we're in 26. And this year we want to launch five more. So the speed that we're able to launch has been completely changed because of stablecoin. And so because that's changed, we can now provide these services to these enterprises in all of LATAM and then eventually want to move to Middle East and Southeast Asia, going back to that vision of a global business bank in 40, 50 countries.

Lance Glinn:

And so let's stick on the Argentina example for a second, because Argentina, obviously, as you just said, is a country with unique challenges and unique currency, and that real currency volatility, as you just spoke to. How does a market like Argentina highlight the real world value of faster and a more flexible global financial infrastructure?

Dileep Thazhmon:

Yeah, because going back to the value of stablecoin, the use case that stablecoin unlocked is that because it's pegged to the dollar, you get access to the dollar even if you're outside of the US. And so when you think about what's the financial unlock for someone in Argentina, it's the fact that now your currency is not being devalued and you can spend that in any part of the world that accepts Visa or MasterCard. So one, you get protection from the currency volatility, but two, you also get acceptance. Which is, you can use that money, you can use that in any place that accepts a card. Previously, Argentina Peso is in Argentina. So you unlock both volatility protection, but then also acceptance and usage across the globe. And you can do that sitting in Argentina. It's a huge, huge structural unlock in how the financial system operates.

Lance Glinn:

And when you build and expand in places with unique challenges, whether it be in Argentina, Mexico or Brazil, name a country, whatever you want, when you operate and expand into those places, has there been any teachings? Have you learned anything specifically of those challenges, of those unique setups that potentially you didn't know prior to going into it that was like a wake-up call? Because I'm sure there were a lot of things that were like, "Oh, we have to do this this way," or, "We have to do this that way." You may have planned something out and then go into the country and realize that your plan needs to go in the garbage and come-

Dileep Thazhmon:

I could go on for hours, but I'll highlight one or two things. So one, it's the importance of localizing. So some products you can sell globally. Cards, you can sell globally. Because if you sell it to an enterprise, what are you selling? You're selling card controls, you're selling the ability to have teams, to have budgets. That's the same thing whether you're an enterprise in country A or country B. Payments is very different, because if you think about payments, you still need access to the local rails. ACH, FedNow, that's just in the US. You go to Mexico, you're using SPEI. You go to Brazil, it's Pix, Boleto, TED. So the big learning was that some countries, like Brazil, you have to localize the payment product for it to get actual adoption. In Brazil, we launched three years ago, the product was similar to the product we had in Mexico, and it just wasn't working on the ground. And so it took us about a year to localize it for Brazil.

And that's what's been driving a lot of the growth, where this year we're tracking to cross 100 million in ARR. And a big driver is localizing the payments that we have for the country. And it took us about a year and a half to realize that without doing that, you don't have access to the local infrastructure. And if you don't have access to local infrastructure as a CFO, you can't do local payments. And if you can't do local payments, that's a very big part of the corporate spend.

Lance Glinn:

And then across these different countries and across North America, Latin America, Europe, Middle East, what have you, are you seeing differences in terms of customer desire and customer needs? Or are you seeing pretty much a basic one or two, "Hey, these are what customers want," regardless of where they are in the world?

Dileep Thazhmon:

Yeah, so we don't sell to consumers. So that takes away a big part of the box. We also-

Lance Glinn:

Well, even for the enterprises that you're selling to, depending on the country that they're in, are there specific things that connect all of them or are they different, depending on if you're in Mexico or Brazil or the United States, or where have you?

Dileep Thazhmon:

Yeah, so it's interesting. When you sell to enterprise, the biggest thing they want is controls. So it's controls for employee spend. Because they want to know what you're spending on, they want to know that you're not, let's not say misusing funds, but you have the ability to spend a card and the card's only for an Uber. So biggest thing is controls. The second-biggest thing is approvals. So for payments, the big unlock that we had was adding a maker checker type system. Before we had that, we could still do payments, but you don't have a flow that works for these companies. So making sure that you have controls, making sure that you have approvals. And then the last one is just covering enough surface area, where you have cards, you have payments, we're just launching AR, that they see you as almost like a single source of a platform that they can use across different needs.

And so for us, once you solve those two, three things, it is fairly scalable. You can sell that product to many different enterprises. Because an enterprise, again, they look the same at a certain size because they all care about the same thing. I care about the same thing as someone else that has a company the same size as me. So yes, there is a lot that you can take from one country to the other.

Lance Glinn:

I want to pivot the conversation to AI for a few minutes. And I know you've described Jeeves as an AI native FinTech, not just in terms of the product, but also how you operate internally. What does, first and foremost, that look like in practice? Because a lot of companies will say, "Hey, we're an AI native company. We operate, we use it internally." And that could be, "Oh, they're just using Copilot to write copy." What does internal AI look like to Jeeves?

Dileep Thazhmon:

And I think it's a great question of what does real AI usage look like? So I'll give you an example of where we see this day-to-day, and then what it's driving. And so if you think about just where we are today, so two years ago we were doing about 400 million in TPV, so that's total platform volume. And this year we just crossed about 3 and a half to 4 billion. And by the end of this year, we'll be hitting about 7 to 8 billion. So it's substantially increased. Now, at the same time, the full-time headcount that we have has actually decreased. So we have less people today than two years ago. Volume has 10xed, revenue has probably 8xed, which has significantly increased, but the headcount is lower. And so that's driving direct impact onto margin. So our actual margin has also increased substantially from 40% to now our gross profit sits north of 80, 83%. So very good structure for a FinTech.

So that's the numbers. Now, when you look at day-to-day, there's a few areas we see direct impact. So one is on the underwriting side. So when you think about underwriting, our core product is a 30-day charge card. So that means we'll give you spend for 30 days, on day 31, you got to pay us back. But that means there's still some underwriting risk for the 30 days. We have to judge whether you're good enough to pay us back. We do that today with a team of four people. So four people are underwriting almost 4 billion in TPP. That just wouldn't be possible before. And that's AI. That's better models. Models that are smarter, they can learn, they actually train themselves. And it's just improving because the actual core model LLM itself is improving significantly. So that's one that we see.

The second one is, if you think about the data that we have, it's very hard to replicate some of the core data that we have. What that means is we have access to data from all these different companies. We can aggregate it and we can train some of the models. So one area we've done this is, if you think about enterprises, they usually have a person whose role it is to look at a transaction that came in and put a GL code to it. And so what that means is, "Okay, this transaction's an Uber, this should go to Taxi," as an example, and then that feeds into your ERP. So we trained an agent that can do that, and it does it with 99.9% accuracy. And we've had about 2 million plus transactions already coded in the last two months.

And so that's another example where that's just actual human work that is now done in five minutes, and it's actually done better than a human. And so it doesn't mean that we've figured it out across the board. I'm still constantly pushing the team to be faster, because again, I think us, saying we're AI native, only is as good as the outcomes showing itself. And to me, even at the size we are today, I think we can get faster and more efficient. And so that's the push that I'm doing with our team.

Lance Glinn:

When you said... And I want to go back to before, you said you have four people looking over, I think it was 4 billion in volume. I don't remember the exact word that you said. When you say that out loud, compared to when you started Jeeves, how surprising is that, that we're there? Because that is really crazy to think about. If we're just talking in human terms, it's essentially a billion for every person, which is unsustainable without the use of technology. But to see that number and to see the limited amount of people you need to hit that-

Dileep Thazhmon:

It's a structural change. That would've been 10 to 15 people before. But the way I encourage the team as well is, it's four humans and then you also have, I don't know, 10 agents. So it's really a team of 14 with humans and agents. But it's just that it's changing so fast. It's actually a really good point, because a lot of times people are like, "Oh, AI isn't doing this." And it's like, "Okay, let's say you had a human join, you would train them. You don't just be like, 'Hey, come on in.'"

Lance Glinn:

Come in and sit down.

Dileep Thazhmon:

And then one hour later, you're like, "Well, you didn't figure this out." It's like, "No, you train them." You invest the time, you work with an intern, you work with an analyst. That's how you have to train AI. You do that, you'll get the outcomes. If you don't train, how can you expect the outcomes? And a lot of people don't do the training, and then it's like, "Hey, AI is not working." It's like, "Did you train them? Then why would you get an outcome?"

Lance Glinn:

How do you personally... You're leading this company, having to focus on all the different parts of it. You can't just focus on one. You have to lead every part from, I'm sure, HR to obviously technology to communications. You're overseeing everything. How do you stay adept with the ever-changing technology, the ever-changing AI? Do you vibe code? Do you create your own AI agents? How do you stay in tune with everything that's changing?

Dileep Thazhmon:

Yeah, so a few thoughts there. So one, a long time ago I used to be an engineer, so I actually feel more unleashed now with AI. Maybe my team doesn't enjoy it as much, but for me, it's been a huge, huge unlock, because I feel like I can actually get in and do stuff. I personally don't believe in a chief AI person. The chief AI person is the founder and the CEO. And if you are not doing it, you shouldn't expect anybody else to do it. So as an example, in our company, I was one of the first to actually build an OpenClaw. Not because my OpenClaw is good, it's pretty crap, but if I do it, then people are like, "Oh, okay, he's doing it. Now I feel like I'm-"

Lance Glinn:

It's a lead by example-

Dileep Thazhmon:

Exactly. Because people are always concerned. Is there going to be security issues? And by the way, there are security issues, but if you don't put your feet in the water, or whatever metaphor you want to use, you're going to be so far behind that it's literally going to limit your career. It's not a Jeeves thing anymore. It's like saying, "I don't know how to use Microsoft Word in 1995." You can't be hired. And that's what I've been trying to push the team a lot. I do think me, personally, and this is why I love Jeeves and what we do, is it's such a high steep learning curve, and I love that. For me, that gets me excited. But it can also be exhausting because you have to cover multiple countries, you need to know regulation, the product keeps changing.

But for me, that learning curve is just so, so important for my day-to-day growth, and AI just accelerates that as well. So that's what I push. And I think the hard part is making sure you hire people that also reflect that same relentlessness and curiosity. Because if you don't, it becomes a part of the org that starts dragging down, and you want it going in one direction.

Lance Glinn:

And you do want it going in one direction. And obviously Jeeves has had a lot of success and is growing exponentially, but how do you make sure that, not just people you're bringing in, but people that have been at the company, don't get complacent? Because I do think, especially with a company that is growing at the speed that Jeeves is, that complacency can come fairly easily. You don't want that to obviously happen. You want to always think there could be higher ground. So how do you make sure that happens?

Dileep Thazhmon:

So we have a value internally, which is what we call relentlessness. And so it's this idea that you have this motor that can go to stage four or five, six. It's a really hard thing to figure out in an interview, but you can probably figure that out in the first two weeks, whether a person has it. And what we try to do is, we try to hire quickly, but also fire quickly if the person's not there. And if it's not there, it's not going to come. And you can tell yourself all the stories you want, that it's going to come over time and the person's going to figure it out, they're not. And you usually know that in the first two weeks. And so we try to sense that out in the interview questions.

I like to ask this question of, "If you were hiring for this role and you saw someone that was good versus someone that's exceptional, how do you see that delta? What's the difference between good and exceptional, as an analyst?" And then that gives you a really good picture of how they think about what's baseline, and then why you need to be excellent at your job. And so there's different ways to suss that out, but you usually need someone actually doing the job. Because the thing I've learned with some companies, like Jeeves, and there are other companies that move quickly as well, is if you can't parallel process, you are going to be paralyzed. Because there's just so much information coming and you have to be able to transfer knowledge from one hand to the other and do multiple different things, or you're just going to drown. And some people love that. And again, that is part of what they enjoy. And for some people that's exhausting. And so that's how I look at it. But yeah, we call it relentlessness inside, and that's to me the number one driver of success.

Lance Glinn:

And I like that line. If you can't parallel process, you're going to be paralyzed. I think that makes a lot of sense. And I think you're 100% right on that. Dileep, as we come closer to the end of our conversation, how do you just see, for the industry as a whole, corporate banking evolving from today's, I guess, local bank-led systems towards something that is more real time, that is more global and potentially more AI driven?

Dileep Thazhmon:

Yeah. So one, obviously, I think banks have an incredible role in the ecosystem. And people always say we're going to replace banks, but I don't think you can replace banks. You have to supplement them. You have to make the offering better for the end user. I do think where we're going, and I think there's still levels to get there, is getting to a form where money is more programmable. So what that means is, to program money, first you need money to be digitized. So just taking an example, like Brazil, 90% of the population use digitized currency. They don't have cash. There's no cash. So in the US, we're still working on that, to get to that point. But once you have digitized currency, then the next step is how do you make that programmable? And what that means is we're still getting to that endpoint, where in the next year, in the next two years, I see a lot of the structures of an org, including finance, including operations, being run through agents. And for agents to be powerful, they need to have money that's programmable in a way that they can use it.

And to me, this is what converges both AI, which is agents, and then stablecoin, which is what takes the money and breaks it up into small little pieces so that it's then programmable. So if I look at what's coming in the future, I see banks as being a catalyst, with FinTechs, hopefully, that can do part of that development cycle for getting to [inaudible 00:30:32] money, but then you need bank adoption for that to actually get to scale. Without that, it's always going to be a niche product.

Lance Glinn:

And so we've talked about how Jeeves has already expanded, already scaled, already grown at a significant rate, and the opportunity for even further expansion in the years to come. What does that next chapter look like for Jeeves? If we're sitting here again five years from now, what are we talking about?

Dileep Thazhmon:

Well, hopefully it'll be ringing a bell here in five years. But for us, there's a few things. There's geographical expansion. So we have covered 25 countries. We just launched Argentina on corporate cards. We want to do the same for all of LATAM, and then expand to Middle East, Southeast Asia. So the one is the geographical box. The second one is product suite. So we cover cards, we also cover AP. We're adding AR, and next is going to be treasury, and so on. So we want to own the whole CFO suite. And then the last one is, for all of these, you need to combine both agents and services. So what we've learned, and this goes back to one of your questions, is even if you have the agents or someone doesn't know how to use them, it's almost like you don't have the agents because nobody's using them.

And so part of what we're testing right now is, we go into a company, map out their entire flow for closing the books for reconciliation, and then build agents, and the human. And then we have the human certify it. And then we talk in the language that the CFO knows, which is, "We will charge you one headcount instead of seven, but get to the same outcome for you." So those are the three lanes that we see, geographical expansion, product expansion, and then AI plus services.

Lance Glinn:

Dileep Thazhmon, CEO and founder of Jeeves. Dileep, thank you so much for joining us Inside the ICE House.

Dileep Thazhmon:

Thanks so much for having me. I really enjoyed the conversation.