Sierra launches Ghostwriter to automate agent creation

Last month, San Francisco-based startup Sierra launched Ghostwriter, an agent designed to build other agents. With this tool, the startup intends to replace traditional click-based web applications with natural language. Users simply describe what they need, prompting Ghostwriter to autonomously create and deploy a specialized agent to execute the task. Speaking this week at the HumanX conference in San Francisco, Bret Taylor, co-founder and CEO of Sierra and former co-CEO of Salesforce, argued that the era of clicking buttons in software is ending in favor of natural language interaction. Taylor noted that many enterprise tools are not used regularly, pointing out that employees typically only sign into systems like Workday when onboarding as a new employee or during open enrollment. Instead of learning to navigate complex systems, he argued that users will soon use natural language to complete tasks without ever interacting with the software interface, stating that he truly thinks that is where the world is going. “Most companies don’t want to make software,” Taylor said. “They want solutions to their problems.”

Sierra is already leveraging Ghostwriter to deploy agents at what Taylor describes as unparalleled speeds. As an example of this rapid deployment, the startup implemented an agent for retailer Nordstrom in just four weeks. This speed supports the company’s rapid financial scaling since its founding. Sierra’s key financial milestones include:

• Reaching a $100 million annual revenue run rate (ARR) in less than 21 months after its founding, a milestone the company announced last fall.
• Securing a $10 billion company valuation.
• Raising a $350 million funding round led by Greenoaks Capital in September.

While a fundamental shift in software may be coming as Taylor predicts, several technologists and investors note that current implementations are far from autonomous. Many companies claiming to offer AI agents, including Sierra and legal AI startup Harvey, employ “forward-deployed” engineers who must constantly update and fine-tune customer agents to ensure they work as intended. These forward-deployed engineers work directly on customer implementations to maintain AI agent performance. This operational reality indicates that the transition to fully autonomous natural language interfaces still relies heavily on human engineering to keep systems functioning correctly.

Why it matters

Sierra is attempting to shift enterprise software interaction from traditional click-based interfaces to natural language-driven autonomous agents, though critics argue the technology currently relies heavily on human intervention.