Ramp Co-Founder and Co-CEO Eric Glyman on AI Adoption, Jobs, and Token Spend
In a Nutshell
Ramp’s data shows that high AI spend correlates with roughly double revenue growth over three years versus 10% for non-adopters, and firms spending ~$30 per employee per month on AI hire 12% more people than peers. Ramp itself is processing 1% of U.S. corporate payments, growing >100% YoY at a $60 B valuation, and delivering 7.5% expense reduction for customers. Token spend is emerging as a third major business cost category—already 2% of white-collar labor spend—and CFOs now face the challenge of managing unpredictable, agentic AI usage while marketers focus on being interesting rather than broadcasting messages.
These notes were generated by AI and may contain inaccuracies.
Many companies have brilliant people but approach marketing by asking how to get their message out there, which kills ideas. The question the team asks is how to be interesting and who is doing interesting things in the world.
Eric Glyman is Co-Founder and Co-CEO of Ramp, a finance automation platform now worth $60 billion. Ramp makes software that helps companies manage expenses. From one place, companies can issue smart cards that know the expense policy and do expenses automatically. Ramp handles payments of all kinds including check, wires, ACH, and automates accounting. The effect is businesses need fewer tools to run their operations and spend less money. A typical business using Ramp cuts their expenses by 7.5% per year.
Most large providers in the corporate card space market spend more to get more. Ramp designed a business model where the more they help customers save in money and time, the faster they would grow. Ramp's net promoter score is in the high 50s, comparable with a company like Apple. The Head of Design at Ramp says they are trying to make a product people love in a space people usually hate.
Ramp went from a $44 billion valuation in June to $60 billion four months later. Ramp is doing well over $2 billion a year in revenue, growing north of 100% a year, and growing 25% faster this year than last year. About 1% of all corporate payments in America flow through Ramp, meaning 99% of payments are not on Ramp. If Ramp doubled again, they would only be at 2%, then 4%, then 8%, showing significant room for growth.
Ramp launched in 2020, a month before the pandemic began. No one was traveling, but it turned into a blessing because everyone needed to figure out how to manage expenses remotely. In 2022, when interest rates spiked, it slowed down and killed many competing businesses. However, Ramp's value proposition of reducing waste resonated more strongly during this period. During 2022 and 2023, Ramp raised more money even at a lower price in a down round, hired more people, and invested in technologies that kept saving people more money, allowing them to pull ahead when others were declining.
Eric's job is to answer the question of how to create the conditions for his team to do their life's work. He constantly works to debug organizations, understand how to speed things up, and get to decisions faster. He looks at metrics such as how fast decisions get made, how long it takes a 24-year-old with a bright idea to buy a new piece of software, how many decisions are being made, and how many lines of code or new features are being shipped. He also looks at the percentage of time the team spends in recurring meetings. At the start of each year, Ramp cancels all meetings and only adds back those that are important.
Ramp sits atop tremendous customer data from over 85,000 businesses. The aggregated data shows what feels like two economies. The biggest change from years ago is around AI adoption. About a year ago with approximately 50,000 businesses in the dataset, Ramp separated AI line items to see if they mattered. The top quartile of businesses with most intense AI spend over a three-year period after ChatGPT came out roughly doubled their revenue. These were not just tech businesses but included a roofer in Utah and a construction firm in Florida, who were growing faster with less overhead. The bottom quartile of businesses that did not adopt AI grew about 10% over the same three-year period.
Ramp puts out a monthly AI index that provides a real-time snapshot of how businesses are adopting AI. Adoption is happening very quickly and at a faster rate than government data suggests. Census estimates show roughly 22% to 25% of businesses are spending on AI, but actual spend data shows significantly higher adoption. The Federal Reserve in the FRED database started pulling in Ramp data directly because it provides a more real-time indicator of business spend. Tech-forward businesses show heavy adoption, as do financial services businesses. Healthcare, restaurants and food service, and manufacturing show 25% to 40% adoption rates.
Current estimates show $160 to $180 billion in spend on inference for AI models. Approximately $8 to $10 trillion a year is spent on white-collar labor in the US, meaning AI inference spend is already about 2% of that total. Some businesses spend more on inference than on their team. About 55% of businesses spend on AI. The median dollar spent per employee is about $12 per employee per month, while in the top 1% it is almost $10,000 per employee per month.
Ramp is growing 25% faster this year than last year at over $2 billion in annual revenue, partly because AI has made the sales team, engineering team, and product team more productive. The data challenges the idea that AI eliminates entry-level work. Ramp partnered with Revelio Labs, which looks at jobs data paired with LinkedIn data and other sources. At high-intensity AI usage of $30 per employee per month, which is like a license or two, firms start hiring more. These firms, despite growing at the same rates, were hiring 12% more people than ones that did not adopt AI or would adopt it later. This was the first quantitative finding showing that spending on AI leads to more job growth.
Token spend refers to what companies spend for AI usage. The last two months have been among the most fascinating in the data. For many years, the narrative was about Jevons paradox, where reducing the price of commodities could increase consumption enough to more than offset the price reduction. This was happening for multiple years with exploding revenues for AI companies. Models are incredible for coding, but many companies using them are paying for more stuff. Individuals may be somewhat more productive, but companies still have bottlenecks. CFOs are now in an ROI era and need to see return on investment. Companies are starting to assess how to use the right model for the right job, not needing an Olympic athlete to walk a dog or renting a Ferrari to pick up groceries.
Token spend represents the third pillar of business spending. The first two pillars are hiring people and hiring vendors. Intelligence came as a third option. Most dollars going to spend are called agentic, where a question gets asked and a session spins up. A computer hires an agent, which may hire another agent, and they debate it out. A question could cost 10 cents, $1, or $20. It is like hiring a contractor to fix something in a house where the bill is unknown in advance. In some sense it is like hiring people, but in other ways it is like a vendor relationship where you do not own the intelligence but rent it from an outside party. Spend on AI grew 21 times over the last year in Ramp's data. Many CFOs lack the core capabilities to manage this third pillar of spend.
The economy is in a surprisingly healthy place. American businesses are more competitive. There are more new businesses now than ever seen before. It is hard to be a small business owner, needing to find a lawyer, someone who knows how to do the books, and someone who knows how to do sales. More small businesses are starting and succeeding than in the past.
Despite buzz that AI would be the death of software, enterprise software seems here to stay. Access at low cost to intelligence will create a lot of goodness and allow people to pursue things they want. Enterprise software helps businesses set governance, storage, and define business processes. The moment with AI models is like hiring a precocious college or high school student who is really smart but does really dumb things on day one. They need to learn the ins and outs of how the business works and why the structure exists.
Ramp has taken a very unconventional approach to marketing, almost like a consumer brand. One campaign put Kevin from The Office, Brian Baumgartner, in a glass office in the Flatiron District for eight hours to do bill pay. It was one of the most exhausting but rewarding days for Brian and became the most viral thing Ramp has ever done, with 300 to 400 million people watching it. The typical marketing question of how to get the message out there kills ideas. Instead, the team asks how to be interesting and who is doing interesting things in the world. Another campaign featured Tony Hawk and Bam Margera on a ramp skateboarding.
Ramp's hiring philosophy focuses on hiring people earlier in their careers and hiring people who spike in certain areas. Job applications often start with requirements that get added to until the description becomes very long, filtering out people and finding someone average across many dimensions rather than exceptional in the one thing that matters most. When looking for spikes, Ramp asks if there was one job this person was doing and one way to measure it, could they find people with extraordinary ability in that area. One of the best hiring pools for finding young engineers is kids aged 14, 15, 16 who played Minecraft and learned to code early, making their own private servers. These are people who have been building great systems for 10 years.
Eric's older brother had learning difficulties growing up. Eric looked up to his older brother as a hero and saw someone who was brilliant when he wanted to be, knowing more about history and having extraordinary research ability. Yet because it was harder for him to write, he would get held back from schools looking for people who were perfectly well-rounded. There were some kinds of problems where his brother was better than anyone Eric knows. The world is filled with people like this. One engineer on the team went to community college, was part of a Minecraft group, has autism, and is able to focus on and obsess over systems, going deeper and deeper in a way most engineers cannot.
When building systems of extraordinary scale, the intelligence of the people involved really matters. There are many people with great talent who are not given a fair shot to show what they can do.
Ramp aspires to one day be a public company. The current focus is on building a business that stands the test of time, generating cash and growing faster.
Tim Coolican is a portfolio operating partner at Blackstone. Before joining Blackstone, Tim spent his career building some of the world's best-known consumer brands, including running L'Oréal Paris in the United States. He works with portfolio companies on brand, marketing, and growth initiatives.
Portfolio companies are spending 20 times what they were a year ago on AI. The pace of adoption is moving at light speed, and the focus is on turning that adoption into productivity and real business results.
Ramp's data shows that AI adoption does not equal job loss. Employment numbers grew by 2.7% across North American portfolio companies, challenging the conventional narrative.
Ramp operates more like a B2C company with marketing stunts and tactics. The brand's why is clear: to save time and money, an eternal need. The product delivers this value in an engaging way.
The first question asked is "how do we be interesting," not how to get the message out. In the current era of discovery and decision-making, it is hard to get and keep people's attention. The focus on getting people to care unlocks creativity, which drives interest, engagement, trust, and loyalty.
The greatest brands are inside out. Ramp's mission and why are clear: to save people time and money. The product delivers on this promise in a unique way, so who they are on the inside is what you see on the outside.
Keep Blackstone in your library
Save the videos and channels worth coming back to, and find them again in one place.
