How Kavak Rebuilt Itself Around AI Agents | Alejandro Maza Ayala
Summary
Angela Strange and Gabriel Vasquez interview Kavak’s Chief Product and AI Officer, Alejandro Ayala, about rebuilding the used-car marketplace as an AI-native company. Ayala explains how Kavak deploys hundreds of thousands of long-running agents daily to handle sales and financing, outperforming humans by focusing on rigorous evaluations rather than superficial adoption. The discussion covers the necessity of top-down organizational transformation, upskilling employees through the "Jedi Academy," and why founders should design new businesses around AI capabilities rather than retrofitting ex…
Topics discussed
Introduction: Kavak's AI-native transformation and agent scale
Alejandro Ayala's background and Kavak's business model
Architecture: One agent per customer with long-term goals
Redesigning the company: From transactional to relational
The critical role of evals in scaling AI agents safely
Superhuman sales agents outperforming human experts
AI in regulated financial services and underwriting
Case studies: AI CEO in Cuernavaca and mechanic sidekicks
Organizational change: Jedi Academy and upskilling employees
New org structure: Flat teams and humans working for agents
Advice for leaders: Top-down strategy and measuring token ROI
Technical pivot: Destroying old workflows for self-improving agents
Creative destruction: Why new AI-native companies will win
Conclusion: Advice for founders in the age of AI
Listen ad-free on Castria