Interview, Fireside Chat
Kavak's Playbook for Rebuilding a Company Around AI
a16zAngela Strange, Gabriel Vasquez, Alejandro Maza Ayala, Ale Massa, Carlos, Gabe, Joseph Schumpeter, Edison, Ford
- Ale Massa projects that by 2035, AI capabilities will reach levels sufficient to perform the duties of a CEO and handle complex, high-stakes financial decisions like underwriting end-to-end.
- The organization anticipates operating with "agent per customer" architectures where between 100,000 and 200,000 agents are instantiated daily to serve 10 million customers, aiming to handle 96% of interactions and 95% of transactions without human intervention.
- Specific performance targets include approving car loans in under three minutes compared to the typical two months, converting customers 2.1 times more effectively than human teams, tripling Net Promoter Scores, and increasing customer lifetime value through agents that remember interactions spanning years.
- The "Jedi Academy" training program is designed to upskill all employees, from mechanics to engineers, enabling them to launch production-ready agents within six weeks, with a continuous update cycle required monthly or bimonthly to maintain relevance.
- Future organizational structures are expected to evolve into flat senior teams where roles are restructured into building agents, working under agents, or performing physical tasks, with human work remaining primarily in domains requiring physical dexterity like the 800 mechanics in Mexico.
- Strategic plans dictate a top-down transformation approach, requiring leaders to define a clear vision for the company's 3-to-5-year future to avoid superficial efficiency gains of 6% to 10% and instead achieve 10x improvements through deep operational redesign.
- Risks include the potential destruction of incumbents that fail to fully redesign their operations, as new AI-native entrants utilizing self-improving loops will disrupt the market with exponential growth in intelligence and economic value.
- Material financial expectations include activating just 1% of the customer base to generate hundreds of millions of dollars, reducing warranty rates by 26% via AI sidekicks for mechanics, and leveraging a self-improving loop where one agent's mistake is propagated to 200,000 others within a day.
- The company expects to measure success through relational metrics focused on lifetime value rather than transactions, utilizing agents that can plan long-term strategies, quote cars, and manage financing, insurance, and car trading-in within a single interaction.
- The vision extends to a "self-improving" organization capable of delivering economic value for the next 4,000 years, driven by access to powerful AI tools for almost free or $20 a month, allowing founders to build rapidly while human and agent collaboration defines the "New World" of work.