Affirm’s AI-native transformation & how it’s driving operational excellence w/ Geddes Munson #259

Affirm’s transformation from a fast-moving fintech to an AI-native engineering powerhouse reveals a rare alignment between culture, technology, and business strategy. At the heart of this shift is a pivotal moment: a three-hour Black Friday outage in 2024 that forced a hard pivot from 'ship fast' to 'ship safely'—a cultural reset that prioritized reliability without sacrificing velocity. This wasn’t just a technical overhaul; it was a company-wide reorientation. The breakthrough came with Opus 4.5, which finally enabled AI agents to write production-grade code. Affirm responded by launching 'AI Retooling Week'—a bold, company-wide experiment where every engineer was given a week off meetings and deadlines to deliver a fully agentic PR. By making the risk collective rather than individual, adoption soared. The result? Over 60% of PRs are now AI-generated, accelerating delivery while simultaneously improving system stability. This isn’t just about automation—it’s about redefining engineering excellence as the ability to scale safely through intelligent agent collaboration. The future? Agentic commerce, where AI acts as a financial advisor embedded in every purchase, personalizing Affirm’s simple-interest model in real time. The lesson? True operational excellence isn’t achieved by adding more work—it’s achieved by removing friction through systems that empower people, not replace them.