Eric Glyman, Co-founder of Ramp

What It Covers

Ramp co-founder Eric Glyman explains how his company serves 70,000+ businesses by treating financial automation as knowledge work, not banking. He covers AI-driven expense management, hiring for high agency, organizational design in the AI era, and why AI labs—not fintech competitors—are Ramp's true competitive benchmark.

Key Questions Answered

• •North Star Metric: Ramp measures success by two numbers only—dollars saved and hours eliminated per customer. The median business on Ramp cuts expenses by 5% annually and grows revenue 16% versus the 3-4% US average. Every product decision is evaluated against whether it moves those two metrics, not feature count or engagement.
• •AI Organizational Restructuring: As LLMs enable determined generalists to perform specialist-level work, companies should question whether traditional departmental silos still make sense. Ramp is actively evaluating its org shape because a skilled generalist using AI can now extend across engineering, sales, and finance boundaries that previously required separate headcount and handoffs between teams.
• •Hiring for Proof of Work: Skip credentials and look for asymmetric evidence of obsession—GitHub activity, leadership in niche communities, or self-built revenue streams. Ramp found engineers who built profitable Minecraft servers at age 15. Two days of working alongside someone yields more signal than 15 hours of interviews, so referrals from people with insider knowledge are prioritized.
• •Token Spend as the Third Budget Category: AI compute spend is becoming a major corporate expense category alongside payroll and vendor contracts. Anthropic and OpenAI are tracking toward $300 billion combined annual revenue—roughly 1% of US GDP. CFOs currently have no structured way to track, attribute, or optimize this spend, which Ramp is building infrastructure to manage.
• •Scoreboard Visibility Drives Accountability: Ramp displays its core customer-savings metrics on physical office walls, in Slack's largest channels, and through queryable internal AI agents. Leaders should define a small number of meaningful measurements—Ken Griffin reduced Citadel's risk exposure significantly by putting a handful of critical metrics on a permanent headquarters display everyone could see daily.