“Companies are going from zero to millions of users and surpassing $100M ARR [annualized recurring revenue] in less than two years—a growth trajectory unheard of before AI,” a16z’s Olivia Moore wrote last week.

One of the best examples of a startup with bonkers growth is corporate credit-card company Ramp. It has been hyperscaling since its inception; it was the fastest New York startup to ever reach billion-dollar unicorn status, hitting that mark within two years of its 2019 launch. At that time, it was on a $10 million revenue run rate (yes, that’s quite a high multiple on revenue). 

One year later, Ramp’s revenue run rate jumped to $100 million. Recently, the startup announced exclusively in Fortune that it had surpassed $1 billion in annualized revenue, not long after achieving a sky-high $22.5 billion valuation in a recent round of financing.

Fortune‘s Leo Schwartz sat down with Ramp’s executive team, investors, and competitors to learn how it has…ramped up…so quickly. The result is Fortune‘s latest cover story. 

I also sat down with Ramp CEO Eric Glyman at Fortune’s Brainstorm Tech conference last week to record a live episode of my Fortune 500 Titans and Disruptors of Industry vodcast. 

During the interview, I asked him what he perceives as the conditions for this era of unprecedented startup growth. I also asked how he has scaled himself as CEO to meet the moment.

Glyman responded that he takes an always learning, always self-improving approach, calling on mentors like OpenAI’s Fidji Simo or Microsoft’s Satya Nadella when he needs advice. Often, he’ll try to mentally put himself out of a job, questioning his priorities and delegating to a more-than-capable team. You can watch the full video interview here or above, and subscribe to future episodes of Titans and Disruptors on Apple or Spotify.

Here’s some of what we discussed in our sit-down:

On Ramp’s explosive growth and valuation:

• How Ramp added more than $6 billion to its valuation in two months
• Why Glyman sees plenty of room for continued growth
• How maintaining a “sense of urgency” helps keep the company moving
• What makes the company irresistible to venture capital firms

On AI:

• How Ramp is using AI to automate tedious expense reports—and why he sees the technology freeing people from monotonous tasks at scale
• How AI is actually helping Ramp’s business as recent studies scrutinize the technology’s efficiency
• How Ramp is using AI to fight AI, particularly in cases of fraud

On readying himself to lead a fast-growing company:

• How Glyman has hyperscaled himself as a CEO, by focusing on his strengths and delegating to employees he trusts
• Why he relies on mentors like former Instacart CEO Fidji Simo and Microsoft CEO Satya Nadella for advice

Listen to the vodcast or read the transcript, which has been lightly edited for length and clarity, below.

Ramp’s explosive growth—how the company got there and how Glyman makes sense of it

Eric, thank you so much for being with us here today and at a big moment in time for Ramp. You are one of the hottest startups—you raised at a $16 billion valuation over the summer and then, like eight weeks later, raised at a $22.5 billion valuation. You just crossed $1 billion in annualized revenue, 45,000 customers.

But first I want to just talk about that number. You look at $1 billion in revenue and then a $22.5 billion valuation. Is the math mathing? Or are we in some valuation hype cycle? What is happening? How does that work?

I think Ramp is just growing so unbelievably quickly. Over the last year, we’ve just about doubled revenue. The fastest-growing public software companies, for reference, expect and hope to grow something like 20% to 30% over the next year. And so the velocity that we are growing at, combined with the scale of the company, is part of what’s getting investors so excited.

But beyond it, I think the unusual part is Ramp is actually growing even faster this year, and doing it while generating more cashflow than we did last year. And so when you combine that with the sheer scale of the market, there’s over $2 trillion spent in the United States on corporate and small-business cards. Which is just one of our markets, and we’re something like 1.5% of that market. It’s hard not to get excited about the potential ahead.

So hyperscale has been in your bones since the company’s pre-launch phase. You and your cofounder, Karim, sat down together and you said, we want to try and create a unicorn company, which is a $1 billion dollar valuation within 18 months. No company in New York had ever done that before.

Why such an ambitious goal? You manifested a billion-dollar company, because you did it within 18 months. And within two years, you had $100 million dollars in revenue run rate.

That’s exactly right. From two years—less than two years from incorporation—Ramp had been valued at not just $1 billion, but $1.5 [billion]. Within two years of the launch of the company, we surpassed $100 million in revenue. And just a few years later, last month, we passed over $1 billion in revenue. For us, I think it’s two things. First, you hit on this aspect of speed. We’re religious about it. We count the days. We’re 2,367 days old.

You know exactly how many days old Ramp is?

We do.

Why?

I think it creates this urgency. I think about leaders like Frank Slootman, who wrote Amp It Up, and just talks about the default state of an organization. Unless someone is driving and leaders are creating tempo, things slow to a halt. The expectation is, you decelerate, and it’s easy to say, you know what? Why not Monday instead of doing it on Friday? We want to instill that urgency to say, today is the only day 2,367 we’re going to have, we’re going to make it count. Also, when every day you’re thinking, What did we get done over the last 30 days? Over the last 60?, you can measure and you can start to make trade-offs and constraints.

You can say, when I look at these last months, these activities really mattered and moved us forward, let’s do more of those. And these other things, even though I liked them, were not as impactful. I have to say no to these things so we can grow faster. And so that’s a big part of it. The last important reason for us is that our whole mission is to help our customers spend less. We want the same for our own company.

What sets Ramp’s corporate credit cards apart

That’s kind of a novel idea, and I want to talk about that, too—the idea for Ramp, and explaining it to make sure everybody understands. It’s flipping the incentive structure on its head in the way that corporate credit cards have traditionally worked, where the more you spend, the more points you get, you’re encouraged to spend more. You actually want people to spend less, which actually seems like a bad business. Is that a business that’s viable?

Well, some of the largest companies in the world are in this line of business. You look at JPMorgan Chase, an over $800 billion company; American Express, a $230 billion company, proving that you can do great by getting people to spend. Now, I sold my last company to Capital One, and I learned how this industry worked, what made it great, but I found it so deeply strange that, at the core, customers were working to make the banks just a little bit worse off by gaming the rewards systems, and the banks were incentivized to go and devalue the reward system to convince people the points were worth a lot and then devalue it in the background. And we just thought, this is a massive opportunity.

What if actually we wanted the same things as our customers, and what if our goal was not to go and give them the minimum points, but actually just help them spend less? You can compete on value. Not competing on price—who’s giving away more? And so I think that was the other motivation in attacking this industry. We believed, and we didn’t know if it would be us, but we thought at the end of the day, this is how the industry should settle. With companies working to make their customers better off and customers genuinely choosing the provider that’s helping them grow. And I think that’s been the big secret behind Ramp’s rapid growth.

So you were not the first startup in this space. There was another competitor, and still is another competitor, Brex, which has a valuation much lower than yours. But it was the first mover, I guess you could say. And at your point of launch, it was already a unicorn. So how have you just plotted along, despite having this big competitor in the space, taking venture capital away potentially, and you’ve just surpassed them frankly in all measures?

Yeah, we were accused a lot in our early days of being the second mover. We always thought we were the 150th mover in this. When you think about companies, most of the juggernauts in this country, they started 175 years ago. Their founders quite literally wore top hats. And so it didn’t bother us so much to come…

You need a top hat.

…we’ll work on it, we’ll talk with the styling team. But look, when we approached this industry, it didn’t bother us to come into this a little bit later. Our view was that this was a large industry that was not aligned with the end customers. And also when your founders maybe wore top hats, I think the importance of time isn’t something you’re thinking about every day. You’ve been around for as long as you’ve been alive, you’ll probably be around…and so what’s the hurry?

We looked at these great companies in the Valley. The Metas, the Ubers, that move fast, that create technology quickly. And it was so at odds with the financial institutions where, if you were transported back in time and had to use the bank accounts or the credit cards of 50 years ago, you’d probably be fine, but if you had to use the phones from 50 years ago, you and I couldn’t do our jobs.

And it just drove home that there was very little product innovation. And so one of the things we set out to do in starting Ramp was, we have got to be first aligned with our customer. [To] help them spend less, be more successful as a business, had to be priority number one. And then number two, we would try to build this valley-type like company that is iterating very quickly, that is measuring in days, that is shipping products every single day. We’ve shipped more products this year than there are business days, more features and announcements.

And the goal when you do that, is the experience of how much time the product saved just expands and compounds faster. And so we’re trying to catch up. What I think the financial services industry should have delivered over the last 50 years, we’re going to try to do it in just a handful, and actually make our customer’s businesses better, because it matters. 

How Ramp is using AI—and if it’s working

You didn’t start out as an AI company, but would you say you’re an AI company now? How are you using it to make Ramp more efficient and your customers more efficient? Is it actually working in a measurable way?

For sure. So first, when you think about our customer base, we support over 45,000 companies of all shapes and sizes, from family farms to the Fortune 500. But for the majority, especially the small- and mid-sized businesses, they don’t have a single engineer at the company, let alone an engineer working to make their finance department modern, adopt AI, all of that. Here at Ramp, we spend over 50% of our payroll on R&D, on engineering, on data science, on design, all focused on integrating the latest and greatest technology. So that even if you’re a small business, you are benefiting from what’s happening in these research labs.

Grace Rivera for Fortune

And so one of the ways that it shows up for a customer is, if you go and you tap a card at the store, you will get a text from Ramp. You snap a photo of the receipt, and we automatically match it to the right transactions. We auto-complete the accounting category. Today, most people are used to expenses being the worst hour of their month. Very painful, takes a lot of work. On Ramp, you snap a photo and you’re done. The entire expense experience takes like 10 seconds.

For most of our customers, they’re not necessarily thinking, I’m buying an AI expense report. It’s just an easier way to do business. And it happens to be that AI is how every single step is being sped up along the process. Does that make sense?

Yeah, it does. And do you feel like the companies are benefiting on the other end from the AI efficiencies you’re able to provide? There are all these studies out—there’s one in particular—that people keep talking about where all these corporate pilots are failing. And actually, people are failing to be able to generate more revenue thanks to AI, more efficiencies from a monetary perspective.

And so I’m curious—has Ramp increased its revenue because of AI, and can you prove that you’re increasing companies’ revenue because of AI?

I love that you asked this question. One of the things that’s very unique in our industry—I think we’re the first, and I still believe that we’re the only industry to actually measure how much money and how much time we have actually saved our customers.

Since inception, we’ve helped our customers spend $10 billion less than they would’ve otherwise spent, and automated 27.5 million hours of work. When you look at the average company though, we actually are able to help companies reduce their expenses by over 5% per year. Compare that to a rewards program. There’s not enough interchange to fund more than the order of two-ish percent of a rebate. We are saving customers dramatically more than what’s possible. And when you look at the history of the company, when you first covered Ramp when we launched in 2020, we thought we could help the average company cut their expenses by 2%.

That’s well over 5% today, in large part because AI is starting to go and complete the expense to do the books and accounting. To go and move money to higher yield. It’s able to not just suggest, but to go and take action as a part of the process. And so I think there are a lot of companies out there selling AI services but aren’t measuring the results, a lot of companies selling you rewards that aren’t thinking about the impact on the bottom line.

Ramp, from the jump, has been focused on: what is the ROI, what is the impact that we’re driving, religious on measuring and reporting that out. And I think that’s part of why our net promoter score is in the sixties. It’s comparable to an Apple, and I think that a lot of companies that are struggling now with all the AI they’ve sold that people aren’t feeling so great about, having the buyer’s remorse, they didn’t start with that simple insight. They should be thinking about: What is the outcome they’re driving, and how do you measure it from the start?

And are you using AI to also fight AI? Because I saw a story the other day about how there are now these AI receipts that look very much like real receipts. And all of our employees are very trustworthy, but there might be a bad egg throwing in some AI receipts in there. Can you catch that? How are you thinking about blocking AI initiatives when it’s harder and harder to prove if something’s real, like an expense?

There’s a variety of ways. First, it was earlier this year when one of the newer GPT-4 models came out, and suddenly it was clear that it was very easy for people to go and generate AI receipts. We partnered with the leading labs—OpenAI, Anthropic, and others—first to create detection systems, but we have a repository of over 100 million receipts that we can look at. We’re using AI to fight AI, to go and block these transactions. It’s something regular systems can’t do.

And next, because we have multiple sources of truth—we have the card and merchant data, we have the image data, we have the receipt data, we have the accounting data—we are much better than single systems, like an Expensify or Concur, where you just get an image and that’s the only thing you have to go on. Because we have multiple sources of identifying whether this transaction occurred, it’s much easier for us to detect what this receipt says, what the amount was, or the way the LLM generated a receipt that looks different than these 1 million other receipts we have for this merchant.

That’s one large way. The second large way—I think a lot of waste happens and fraud happens because managers are too busy. When you take a 100,000-person organization, a lot of people are spending time, probably in this audience, going and checking for your employee, should I approve or deny this expense? But the reality is, you’re busy, you have another job, you’ve probably just hit approve.

We’ve trained large language models to actually read your policy in depth—it probably has read it better than anyone in this room. It’s audited and seen every expense, and we are able, our policy agents are able, to actually go and automatically approve 90% of transactions from the jump. Five percent to 10% that need attention, we can show you why it was in or out of policy. It’s 99% accurate, which is about 10 times more accurate than the average employee. And what it means is, it’s a massive time saver. It’s saving managers from the time of reviews, but it’s also catching a lot of things that people would not catch. People spending company money that, in the old world, would’ve just gone through, because no one had the time to look at it.

And as you’re building all these tools that are AI capable—efficiency and time and money saving can also equate, in a worker’s mind, to, Is that my job you’re coming for, Eric? So I’m curious how you’re thinking about, in the most honest way, the bigger vision: If Ramp is really successful in saving companies time and money, what will that do to traditional business functions? Do CEOs need a whole finance department if all goes to plan? Do they need a human resources department? Eventually a lot of the core business functions operations. Is that the grand vision? 

I don’t believe that AI is smart enough to do the job of a CFO or a complete finance function, but it is definitely capable of doing your expense reports. It is definitely capable of categorizing transactions. And I think for most people, I don’t think you’re adding deep human intelligence when you’re going and snapping a photo and you’re describing what you bought and you’re going and tagging transactions. It’s very low-level work and, for most people, it is just the worst hour of your month. Why not automate these terrible parts of your job away? It allows your best salespeople to go and spend that last hour selling and actually doing the work they were meant to do. And so we’re very much in that phase of creating a lot of delight and joy for people in their roles.

I think when you abstract it and you look more long term, you think about: What is the finance function? Where are people spending time? And at least on the spend side, a lot of it’s really just algorithms. It’s going and determining who should spend what under what circumstances. Once the spend has occurred, how do I categorize it correctly? That takes a lot of work.

And then based on what happens, how do I goal-seek to a better outcome the next time? So much of the finance function today, I would argue, on the order of 80% of it, is actually looking backwards. It’s trying to figure out: What did we do? What did we spend on? What’s happening in the business? It’s not asking the interesting questions that most people in finance got in it to do, which is, How do I make this business better? How do we spend on the things that matter? Where is value? How do I allocate capital better? And I really am a firm believer that the low-level work that people don’t want to do will go away.

But I believe, and I’m fairly optimistic, that when your books are keeping themselves, money finds its way to higher yield. One, for businesses, you’re going to have a lot more at the end of the day. For the average American business, they have an 8% profit margin. If you can go and grow it even by 1%, it’s equivalent mathematically to a 12% increase in revenue. And so I think that bottom-line impact—to create more margin, to invest more—is going to be profound. And second, I think for people, the work is going to be more interesting. At least as far ahead as I can see and imagine, but we’re just excited to be working on it.

The influx of attention, and cash, from investors

What’s it like to be the hot girl on campus? How frothy is it out there, and were you surprised by some of the investor behavior you’ve seen, given your last company only raised $2 million and now you’ve raised over a billion? Slightly different. So, what’s it like out there to be a fundraising startup that every investor seems to want to have a piece of? 

I think for investors, I empathize certainly in the venture industry. There are more investors than ever. 

Everyone’s a VC. 

It seems like it. There is a lot of capital, and I think people are looking to find yield. And some of this speaks to how the world is changing faster than ever. We are in a world now where computers can see and hear and think and reason, and that’s bizarre and has all sorts of profound implications. And I think we are, in some sense, multi-trillion dollar jump balls in lots of industries. And I think that the stakes are very high, and that’s part of why people are looking to invest. I’d also say that companies are growing faster than they ever have before.

Noam Galai—Getty Images

Is that because there’s so much money sloshing around? Why is now the moment? The numbers you’re hitting seem unfathomable from even a few years ago.

One, I think that AI is making people more productive. But two, I just think that when companies are able to grow, and Ramp is doing this while generating cash at an unprecedented scale, VCs look at this and say, how could I not invest in it? Because if you’re doubling each year at this kind of scale, within months, that round that looked expensive, proved to be cheap and inexpensive. And so I think that’s part of what’s driving this demand.

There are fewer companies that are growing faster than ever. But I think about another company, Cognition. It’s a wonderful company that started on Ramp. Cursor is another one. These organizations are not yet two years old but are doing nine figures of revenue. And part of this is, they are capturing the moment and selling new types of services. But the other part of it is, their finance teams are benefiting from incredible technology that, in the old world, it just would’ve been much tougher to build up the skills inside of the company to deal with this growth. And so I just think the tools for builders are better now than ever before.

Does it ever make you nervous to be like, I started this company 2,300-whatever days ago, and we’re worth 22.5 billion? The fulfilling on that, and especially if an IPO is on the horizon and you’re going to be answering to investors… anxiety, excitement?

Look, I’m in my mid-thirties. I think you always look up to people, many in this room who’ve been building great organizations, and wanted to be that one day. And so I feel very lucky to have the opportunity to do this and to be able to work on something that I’m really passionate about. But for me, I think valuations in some sense are a derivative. It’s not the thing, it’s not the reason. Revenue comes from customers genuinely feeling that their trust was well earned. That when they signed up for a product, it actually delivered, and it delivered so much that they told other businesses about it. That we made their business better and more profitable, that they’re able to grow faster.

And in some sense, I think for anyone building the business, you start these things, I believe, because you hope to make a difference in the world in some kind of a way. So the valuation is one thing, but the numbers I care much more about are really: How much did we save customers this month? Did we make people better off? And I think that’s why some of the best engineers in the world want to come to Ramp. I think that’s some of why the best designers are working on … you wouldn’t think that these people are interested in corporate cards and expense management. 

Not so sexy of an industry, but yet you’re crafting great talent.

We think it is now. And it’s not just the hot yellow that the Ramp brand is doing, and the fun ads. I think it’s for people who want to matter in the world and have some kind of an impact. I think this is a real way to do this, and do it quickly.

So Eric, for a final question, I want to kind of get inside of your brain as a CEO. It’s really hard to be a CEO these days, as you know, and navigate all the change. And I can’t imagine what it’s like to go from you sitting there with Karim, thinking you’re going to start this big awesome company, just 2,000-plus days ago, to what you’ve achieved today. How have you scaled yourself? How have you gotten yourself ready to meet the moment of what Ramp is today?

I try to approach it with a lot of humility. There’s a lot of things I don’t know. And I think one of the problems of compounding growth is that, what allowed you to grow by 100% over the last year will, by definition, if you don’t do something about it, you might only grow 50% the next year, 25% the next. And so you can know certainly what got you here will not get you there. And so it forces you to constantly look in the mirror and say, Okay, what was I great at that I need to give up? Because the game has changed a lot. And so I think it’s a lot of just being real about that. It’s not about getting a little bit better at the small set of things, but actually trying to put yourself out of the job very, very often.

Do you mentally try and put yourself out of a job?

I do.

How do you do that? Do you think about what bad Eric could do today? How do you think about that?

Well, there are things that you learn about yourself. For example, I will put it this way. If there are 100 things to do, I’m the kind of person that’s like, What are the top 10 most interesting things? And I’ll do those and drop the other 90. And in the early days, no big deal, but at some point that will kill you, because those other 90 things need to get done.

So I try to look for great operators, people who are not going to drop the ball, people who are better at sales, better at pieces of marketing, better at engineering. I actually think it’s a joy to go and find people who can teach you things, put them into roles, and give them the work. And try to focus on the areas that just I can do, or maybe I have a little bit of an edge, and actually make sure the return to my time is higher.

And so some of it’s trying to surround yourselves with great mentors. I think about people like Fidji Simo. She was the CEO of Instacart, took them public, now she’s at OpenAI. Satya Nadella is a great mentor. And I think some people pursue coaches. I try to call people up for an hour at a time, where if I can just get their advice on AI or marketing or sales and learn just a little bit. Ask them who they’ve learned a lot from in particular fields and just jump from person to person.

And that’s been very helpful. And then last, I think at the end of the day, all a company is is a collection of people. You forget it along the way, but it’s still true. And I think that if you can go and build a strong team, try to empower people to double down on what makes them great, not fix their deficiencies, that’ll help you have a much more well-rounded company. And so I’m still learning. Open to advice and trying our best, but it’s been a very fun ride.

Well, Eric, it has been so fun to watch what you’ve built at Ramp, and we’re going to continue to watch it at Fortune. Pick up the next issue, you’ll see a big feature on Ramp and their explosive growth. But thank you for spending time with us today. 

Thanks so much, Alyson.