August 6, 2025
5 min read
Lee Chong Ming
Vercel's CEO says AI agents, not humans, are becoming software's primary users, transforming APIs and developer tools.
An Accel-backed startup CEO says your next user isn't human — and it's changing how software gets built
Vercel's CEO Guillermo Rauch believes the future of software development is shifting from human users to AI agents. These AI-driven entities are reshaping how APIs and developer tools are designed and built. The future of software isn't being built for people — it's being built for machines, said Vercel's CEO, Guillermo Rauch. "Your customer is no longer the developer," Rauch said on the "Sequoia Capital" podcast. "Your customer is the agent that the developer or non-developer is wielding." Vercel, a web infrastructure startup valued at $3.25 billion last year, is witnessing a fundamental shift: code is increasingly written not just for humans to read or interact with but for AI agents to understand, use, and extend. "That is actually a pretty significant change," Rauch explained. "Is there something that I could change about that API that actually favors the LLM being the, quote-unquote, entity or user of this API?" This AI-first era demands that software tools evolve based on how large language models (LLMs) interact with them. "LLMs' strengths and weaknesses will inform the development of runtimes, languages, type checkers, and frameworks of the future," Rauch said. Rauch also highlighted that Vercel's newer users — including designers, marketers, and AI agents — expect software to "just work." Unlike developers accustomed to debugging and handling errors, today's users have little tolerance for failures. "You want something that works 99.99% of the time," Rauch added, calling this an "amazing pressure" for product builders. Last year, Vercel raised $250 million in a Series E round led by Accel, with investors including Tiger Global and GV.Rise of AI agents
The year 2025 has been dubbed the year of AI agents, which could fundamentally change how the internet works and how users interact with apps and software. Bernstein analysts noted in February that while websites and apps won't disappear, users may no longer interact with them directly. Instead, AI assistants will aggregate information, content, and widgets, becoming "the aggregator of the aggregators." "If it scales and plays out like we think it might, this. Changes. Everything. The aggregators get disaggregated, and much of consumer internet may be structural shorts. Welcome to the Agentic AI era," the analysts wrote. However, AI agents are not without flaws. Researchers warn that agent errors are prevalent and compound with each step they take. Patronus AI, a startup that helps companies evaluate and optimize AI technology, explained that an agent with a 1% error rate per step can accumulate to a 63% chance of error by the 100th step. "An error at any step can derail the entire task. The more steps involved, the higher the chance something goes wrong by the end," Patronus AI wrote. Despite these challenges, guardrails such as filters, rules, and tools to identify and remove inaccurate content can help reduce error rates. Small improvements can yield outsized reductions in error probability.Source: Originally published at Business Insider on August 6, 2025.