Cheeky Pint Podcast Summary

Episode Title: Ramp founder Eric Glyman on the many ways AI is changing corporate spending
Date: February 17, 2026
Host: John Collison (Stripe co-founder, going by "Kevin")
Guests: Eric Glyman (CEO and co-founder, Ramp), Alex (Board member at Wise; active participant)

Overview

This episode features Eric Glyman, co-founder and CEO of Ramp, discussing how artificial intelligence is transforming corporate spending and operational efficiency. Ramp, founded in 2019, has rapidly evolved beyond just corporate cards to become a comprehensive finance automation platform. The conversation dives into AI’s real impact on mundane business processes, the cultural and behavioral psychology of expense management, network effects and moats in software, and the future of both financial infrastructure and work itself. There’s also an engaging side discussion on data moats, SaaS competition, and the enduring influence of Capital One in fintech talent.

Key Topics & Insights

1. Ramp’s Business Evolution & Scale

• Origins and Growth ([00:21]–[02:59])

  • Card-based revenue remains the largest segment, but new products—including bill pay, procurement, treasury, and travel—are scaling, with non-card lines soon to comprise the majority of Ramp’s business.
  • Ramp is helping customers cut expenses by ~5% per year; Ramp customer revenue growth (~16%) far exceeds the US business average (~5%).

    “The average customer last year on Ramp grew their revenue by, I think it was 16%; which compared to in the United States, the average business… is 5%. In some sense what we’re trying to do is be not just a better platform, but a better almost kind of digital brain for organizations.”
    — Eric Glyman ([01:09])

• Diversification of Entry Points ([03:16]–[04:27])

  • Not just about spend cards anymore—thousands of businesses now start with bill pay, accounting, or treasury products.
  • Procurement is cited as the fastest-growing line.

    “There were over the last quarter thousands of businesses that came in just for bill payments… probably the fastest growing line of business is procurement.”
    — Eric Glyman ([03:37])

2. Expense Policy: AI, Culture & Context

• Testing Expense Policies with Data ([04:27]–[07:52])

  • Ramp can correlate strictness of expense policies with growth and efficiency metrics—a kind of back-testing at scale.

    “You can do a correlational study between expense policies and company growth.”
    — Kevin ([05:35])

  • High-growth companies typically use permissive, “trust but verify” approaches (a la Netflix).

  • AI/LLMs can now automate detailed, context-aware policy enforcement that was tedious or impractical to do by hand.

    “Now you can have an agent, functionally, that does that… Today it’s over 99% accurate, which turns out it’s much more accurate than people are.”
    — Eric Glyman ([06:37])

• Behavioral/Cultural Dimension ([08:12]–[11:03])

  • Expenses reflect company culture; transparency can instill a “moral code.”
  • The right policy doesn’t boil down to a simple “if this, then that”—context (e.g., closing a big deal) matters.

    “You want a moral code. Like that’s what you want… how do you actually instill that?”
    — Alex ([08:34]) “I would argue in favor of tools that have more context, that do more jobs of work. Because it… helps you get to what actually is right for shareholders.”
    — Eric Glyman ([09:39])

3. Bill Pay, Payment Network Challenges, and the Slowness of Upgrades

• Antiquated Bill Pay Systems ([11:07]–[15:26])

  • Despite digital upgrades to checks and card rails, bill pay remains stubbornly manual: PDF invoices, bank details sent over email.

    “It’s kind of putting a little bit of lipstick on the pig of the whole system being super antiquated.”
    — Kevin ([12:02])

• Adversarial Timing in AP/AR ([14:10]–[15:10])

  • AR (accounts receivable) and AP (accounts payable) are adversarial: buyers delay payments, sellers want cash ASAP. Check timing games persist because of embedded incentives.

• Data and Billing Innovations

  • Vision: A universal “DNS” for company bank accounts to prevent errors and fraud.

    “You should look them up in some central clearinghouse and that way you confirm you’re not being phished.”
    — Kevin ([13:04])

4. AI & the New Software Engineering Equilibrium

• Blurring Traditional Roles ([16:52]–[18:14])

  • Designers, support agents, and marketers can now ship code thanks to AI tools. Cycle times for changes or fixes are shrinking fast.

    “The half-life of seeing a problem to fixing it is shrinking immensely.”
    — Eric Glyman ([17:19])

• Tech Debt & Code Evolution ([18:14]–[21:57])

  • AI may cause a short-term rise in “spaghetti code,” but could later support self-rewriting, outcome-driven architectures. Mission-critical (e.g., 4-nines reliability) code will remain hands-on.

• Software Company Moats: Data, “Dark Matter”, and Network Effects ([22:08]–[26:08])

  • Proprietary, hard-to-replicate data remains a powerful moat (examples: Velex’s legal records, DomainTools’ WHOIS logs).

    “You have a country of geniuses, data center… if you’re not building towards classic four moats, life gets a lot harder.”
    — Eric Glyman ([21:57])

  • “Dark matter” (unseen, hard-won knowledge from accumulated edge cases and complex operational learnings) shapes durable advantage.

    “There are like 9 million edge cases… The tech debt sounds bad… but actually it’s good because you’ve uncovered every single problem that can go wrong.”
    — Alex ([26:07])

5. The “SaaS Apocalypse” & Competitive Dynamics

• How Disruptible Are SaaS Apps? ([28:59]–[32:13])
  • Some features masquerade as companies and can be replicated; others are sticky due to entrenched processes and complexity (“hostages, not customers”).
  • Moats often come from deep integration, compliance, or aggregated buying power (as in Workday or enterprise HR/payroll).

    “Best companies have hostages, not customers, at least in enterprise software… you have to be in this Goldilocks zone.”
    — Alex ([29:02])

6. Ramp’s Long-Term Differentiation: Selling Time, Not Money

• Platform Scale & Aggregated Leverage ([47:18]–[53:37])

  • Ramp’s unique selling point is saving customers time first, not just saving money—contrasting with traditional financial institutions.

    “They sell money and we sell time… We will sell your expenses. Done.”
    — Eric Glyman ([48:39])

  • Platform ability to aggregate purchasing power could allow for collective price negotiations or new models (GPOs, merchant prepayment/balances, etc.), though Glyman sees biggest leverage in continued process automation.

  • Still immense room for growth, with only ~2% of US business card transactions on Ramp.

7. Macro Impacts: AI, Efficiency, and Economic Signals

• AI Adoption Is Underreported, but Actually Ubiquitous ([33:40]–[36:24])

  • Ramp’s data shows most businesses are already paying for or using AI (ChatGPT, Anthropic, etc.), contradicting official surveys.

• Real-Time Spend Data Reveals Underlying Efficiency Boom

  • US GDP growth acceleration is tied to business tool adoption; companies leveraging optimization platforms like Ramp materially outperform averages.

• Time as the Most Undervalued Resource ([36:44]–[40:14])

  • Typical underappreciation of time savings has a greater impact than line-item cost savings.

    “Human time is incredibly expensive. I think the lean companies got this right…”
    — Kevin ([37:20])

• Unique Vendor Data Unlocks Savings ([39:55]–[41:53])

  • Ramp can alert clients when they’re overpaying for SaaS, provide real-time market benchmarks before purchases, and offer personalized procurement advice.

8. Platforms as Price Arbiters & Demand Aggregators

• Can Ramp Broker Bulk Discounts Like Costco? ([41:53]–[44:40])

  • Group-purchasing organizations (GPOs) and demand aggregation can create negotiating leverage—an area Ramp is exploring.

    “Can you go and say, across the ramp buyer base… this is how many dollars will go towards you; can we negotiate a discount the other way?”
    — Eric Glyman ([42:44])

• Merchant-Specific Balances & Loyalty ([44:40]–[46:27])

  • Concept of businesses prepaying for spend (similar to gift cards), benefiting both buyers and suppliers.

9. Lessons from Capital One

([57:07]–[66:24])

• Talent and Experimental DNA

  • Capital One was built by hiring for raw intellect, not banking background; became the McKinsey of credit for risk and analytics talent, seeding modern fintech.
  • Their foundational insight: differentiated risk segmentation expanded credit beyond the elite, leading to massive growth.

• First Principles and Iterated Experimentation

  • Capital One’s rigorous, data-driven approach and willingness to experiment gave it a lasting edge.

10. The Future of Treasury, Banking, and Company Money

([66:31]–[71:08])

• Businesses get very little yield on idle cash—Ramp’s treasury solution offers much better returns by auto-sweeping funds where they earn more.

• Glyman predicts the market forces will pry some of these profits from traditional banks as switching/optimization tools lower the friction.

  “New businesses… this is not a monopoly… I think that that rate will go up… The easier it is [to move or optimize], the more competitive the rate will become.”
  — Eric Glyman ([67:59])

• Automation will increasingly put capital to work intelligently, making “idle” money a relic.

  “If the dollars in your company have some level of intelligence… there’s some ability to opine where should the next marginal dollar go.”
  — Eric Glyman ([69:11])

Notable Quotes & Memorable Moments

• “The average customer last year on Ramp grew their revenue by, I think it was 16%; which compared to in the United States, the average business… is 5%.” — Eric Glyman ([01:09])
• “You can do a correlational study between expense policies and company growth.” — Kevin ([05:35])
• “Now you can have an agent, functionally, that does that… Today it’s over 99% accurate, which turns out it’s much more accurate than people are.” — Eric Glyman ([06:37])
• “What you want is a moral code… how do you actually instill that?” — Alex ([08:34])
• “Proprietary data is a moat… a lot of times the data is free… DomainTools runs a cron job every day on every website, they do a Whois lookup… all that data is free.” — Alex ([24:47])
• “There are like 9 million edge cases… tech debt sounds bad because it’s a pejorative, but actually it’s good because you’ve uncovered every single problem that can go wrong.” — Alex ([26:07])
• “They sell money and we sell time… We will sell your expenses. Done.” — Eric Glyman ([48:39])
• “The marginal cost of arguing has gone down to zero.” — Alex ([53:59])
• “Most of our customers don’t have a single software engineer… If you can deliver some type of knowledge work [via software], that is immense high leverage value.” — Eric Glyman ([55:11])

Timestamps for Key Segments

• Ramp’s Business Model & Evolution: [00:21]–[04:27]
• Expense Policy, AI, and Company Culture: [04:27]–[11:03]
• Bill Pay’s Resistance to Modernization: [11:07]–[15:26]
• AI’s Impact on Software Teams and Code: [16:52]–[26:08]
• SaaS Competition & Sticky Moats: [28:59]–[32:13]
• Impact of AI on the Economy & Spend Data: [33:13]–[36:24]
• Time vs. Money as a Product: [47:18]–[53:37]
• Platform Leverage & Future of Financial Products: [41:53]–[53:46]
• Capital One History and Fintech Talent Pipeline: [57:07]–[66:24]
• The Future of Treasury & Where Companies Keep Their Money: [66:31]–[71:08]

Final Thoughts

This candid, fast-moving dialogue offers a unique window into how a top founder sees the future of business operations, the deep integration of AI, changing cultural norms around expense and time, and why speed, data, and context (not just features) are becoming the ultimate differentiators in fintech. The episode is both a practical guide for operators and a thought-provoking look at evolving competitive advantage in the AI-and-automation-powered economy.