reSee.it Podcast Summary
In this episode of Uncapped, the host and Bret Taylor explore how artificial intelligence is reshaping software strategy, incentives, and the core architecture of modern enterprises. They discuss the idea that the traditional “systems of record”—databases and the associated workflows—will coexist with AI agents, but the relative value may shift from the database itself to the agents that operate on top of it.
The conversation traces how early software platforms built defensibility through network effects, ecosystems, and high switching costs, and then asks what happens when AI agents can perform many tasks that used to require manual interaction with ERP, CRM, or IT service management systems. Taylor argues that the strength of incumbents may erode as agents become capable of handling onboarding, lead generation, quoting, and other familiar processes, while incumbents still hold some advantages in scale, integration, and existing ecosystems. A central question is whether the role of a system of record will diminish if AI agents handle most tasks invisibly, and how to balance the gravity of the database with the gravity of autonomous agents operating around it. The dialogue suggests that the market will favor platforms and ecosystems that can assemble robust agent networks and offer industrial-grade reliability, especially in regulated industries like healthcare and banking, where compliance and risk management matter deeply.
The discussion then moves to pricing models, with a strong emphasis on outcomes-based pricing over token- or input-based schemes. Taylor explains why tying value to measurable business outcomes—such as successful sales conversions or satisfactory customer support—offers a clearer alignment with customer needs than charging by token usage. They also reflect on the practical realities of making AI work at scale, including edge cases in voice and multilingual support, and the need for teams committed to rapid, reliable deployment that can still navigate complex change management.
The interview ends on reflections about the future of work in AI-centric software, the potential for smaller, intense teams to win in certain markets, and the importance of combining deep domain knowledge with AI fluency to deliver durable customer value. Throughout, the emphasis remains on building products and partnerships that can move quickly, but with a maturity that matches the demands of large organizations and regulated industries.