Bret Taylor — once co-CEO of Salesforce and now the driving force behind Sierra — believes a fundamental shift is on the horizon in how we interact with digital tools. Sierra, his latest venture, is carving out a niche at the intersection of AI and enterprise services, shaping the way large organizations deal with customer support. But it’s not just about faster service—it’s about the entire fabric of digital workflows changing.
Recently, Sierra pulled back the curtain on Ghostwriter. It’s not your average AI chatbot. Ghostwriter is designed to spin up entirely new AI “agents” on demand. To put it plainly: instead of grinding through menus and endless drop-downs in clunky enterprise tools, users just describe what they want in plain English. Ghostwriter takes the request and, without human engineers intervening, builds a tailored AI program—a micro-expert—that handles the task.
Taylor, speaking at the bustling HumanX conference in San Francisco, summed it up with an everyday example. “Most people don’t log into business apps unless they absolutely have to,” he pointed out. “You touch Workday maybe once a year during open enrollment. Navigating it is a pain.” He envisions a near-future where employees won’t need to learn clunky systems. They’ll just say what needs doing, in their own words, and the technology will get out of their way.
There was a certain energy in his voice as he described this coming transformation. “We’re about to see the end of the old interface model,” he said, leaning into the microphone. “Natural language will be the new UI.” To him, the death of the login screen and the rise of ‘describe, don’t click’ is not a distant dream—it’s inevitable.
Sierra isn’t just full of ideas. They are delivering, and fast. Ghostwriter allowed Sierra to craft and deploy a bespoke agent for retailer Nordstrom in barely a month—an impressive pace in a world where customizing enterprise software often drags on for quarters.
And the market has noticed. Sierra hit a $100 million annual revenue run rate less than two years post-launch and was recently valued at a staggering $10 billion, buoyed by a $350 million funding round led by Greenoaks Capital.
But there’s a reality check, too. As Taylor spoke about the coming revolution, skeptics in the tech community quietly noted that “autonomous” AI agents are still more ambition than fact. Many startups touting end-to-end AI do, in reality, rely on teams of “forward-deployed” engineers. These human experts are frequently required to adjust and reprogram the AI, ensuring it actually works for clients in messy real-world conditions.
Sierra is no exception, nor is Harvey, a legal tech startup specializing in AI-driven contracts. For now, fully self-sufficient AI agents remain a work in progress, more like high-performance race cars that still need a pit crew, than magic robots you can just set loose.
Outside the Sierra story, the tech scene in San Francisco buzzes with anticipation about where all this is going. Event banners for massive gatherings like TechCrunch Disrupt 2026 are scattered across the city. Industry leaders, VC heavyweights, and hungry founders will soon pack into cavernous halls—vying for the next big thing, trading insights, eyeing breakout startups, and, of course, chasing the elusive edge that might shape the next era of tech.
None of this—the capital, the hype, the hunger—changes Taylor’s main assertion: companies don’t actually want software. They want solutions, solutions that melt into the background, quiet and effective. “Nobody cares about the software itself,” he told the conference crowd. “They care about results.”
He’s convinced the next wave isn’t just about newer apps. It’s about eliminating friction entirely. A world where the gap between intention and action shrinks to a phrase, a sentence, something spoken aloud—and your need is quietly, efficiently met by invisible digital hands. Sierra, and others like it, are betting everything on that future arriving much sooner than we all expect.
