Interview, Fireside Chat
Using AI to Build “Self-Driving Money” ft Ramp CEO Eric Glyman
- Ramp's core mission is defined as creating a "command and control" system for company finances to save time and money, resulting in an average of 5% annual expense savings per company and billions in total savings over five years.
- The company currently serves over 25,000 businesses, ranging from startups to large entities like Shopify and Virgin Voyages, saving the equivalent of thousands of years of labor.
- Eric Gleiman, co-founder and CEO, distinguishes Ramp from typical AI hype by rejecting chatbots as the primary interface for finance, arguing instead for "zero-touch automation" that operates invisibly in the background, akin to "self-driving money."
- Gleiman traces the concept of AI agents back to 2015 with Paribus, a tool that automatically scanned emails to claim price-drop refunds, noting that while agentic AI feels new, the primitive of software acting on a user's behalf has existed for a decade.
- Ramp's current AI strategy focuses on eliminating the "two-app" friction (e.g., separate Amex and Concur apps) by automatically pulling receipts, classifying vendors, and auto-completing expense reports using Large Language Models (LLMs) to analyze transaction data and general ledgers.
- The company is exploring two primary interaction paradigms: "zero-touch AI" for routine tasks that are fully automated, and "agentic AI" where users prompt for specific outcomes, potentially monitored by a "tour guide" interface that guides users through steps or executes actions like card issuance.
- Ramp engineers are required to speak directly with customers upon product launches and remain accountable for performance metrics, a cultural practice Gleiman cites as essential for developing products with superior design and efficiency.
- The leadership views the 2024 shift in foundation models as a "step function" change rather than gradual evolution, enabling human-level reasoning and offering a new durability moat beyond traditional software lock-in.
- Gleiman predicts that in five years, finance leaders will shift from repetitive data entry and transaction tagging to high-value strategic work, though he anticipates that the total volume of work will increase as people move to higher levels of abstraction rather than reducing hours.
- The company's design philosophy prioritizes timeless problems (saving time, money, and effort) over timely tech trends, advising founders to identify enduring customer pains before applying new technology solutions.
- In rapid-fire predictions, Gleiman identifies healthcare (specifically continuous monitoring and diagnosis) and the broader design ecosystem as the next major industries to undergo "self-driving" transformations.
- He notes that traditional financial institutions are hindered by legacy infrastructure and regulatory history, suggesting that true financial services innovation will likely come from non-bank disruptors like Apple and productivity-focused fintechs rather than established banks.
- Gleiman cites Satya Nadella as the most admired figure in AI for successfully transitioning from research to production and fostering a long-term, decades-spanning vision for agentic AI.