As corporations around the world race to integrate and capitalize on artificial intelligence, Vercel believes it has discovered a novel and pragmatic way to stay ahead of the competition: by teaching intelligent systems to emulate the methods, habits, and problem-solving instincts of its most exceptional employees. The result, according to the company, has been a dramatic shift in efficiency and productivity—one that could signal a new paradigm in how human expertise and machine capability converge within the workplace.
Vercel, a cloud-based developer platform currently valued at around $9.3 billion, was founded in 2015 by software engineer Guillermo Rauch. The company’s primary mission is to provide developers with a streamlined environment for constructing, deploying, and scaling modern websites and web applications. Now, however, Vercel has turned its attention inward, employing sophisticated AI agents to handle the repetitive and procedural aspects of many entry-level positions. In doing so, it has managed to shrink what was once a ten-person team down to a remarkably lean operation consisting of one human employee assisted by an automated agent. This restructuring has freed up human talent to focus on higher-order, creative, and strategic work—an embodiment of the company’s belief that AI should amplify human potential rather than diminish it.
Within the realm of artificial intelligence, agents are generally described as digital entities capable of making decisions and completing tasks autonomously. They are designed to internalize instructions, break large problems into manageable components, and independently execute a series of actions—often without direct input from a human user after initial setup. As Jeanne DeWitt Grosser, Vercel’s Chief Operating Officer, explained in an interview with Business Insider, any business process that can be systematically recorded and documented can, in many cases, be replicated by an AI agent. The prerequisite for automation, in her words, is simply the clarity of a well-defined workflow.
Vercel’s journey toward developing its first AI agent began in June with an internal experiment initiated in the sales department. Grosser—who had joined the company only a few months earlier, in March—assembled a small team of three engineers to work on designing agents that could not only reproduce essential sales functions but also refine them through automation. At the time, the company employed ten sales development representatives responsible for managing inbound leads—an essential yet predominantly entry-level set of responsibilities. Among those ten, one individual consistently outperformed the rest.
To create a digital analogue of this top salesperson, the engineers spent six weeks observing the employee’s workflow in great detail. They documented every habitual action, from how messages were prioritized to the subtle tone adjustments used when communicating with potential clients. Using this comprehensive dataset, the team engineered an AI agent capable of mirroring the employee’s process and eventually executing it independently. The result was a digital assistant able to perform the same high-quality tasks at scale, around the clock, and without fatigue.
According to Grosser, Vercel’s newly created “lead agent” now performs tasks that were once distributed across multiple human sales representatives. Using advanced tools, including OpenAI’s Deep Research system, the agent can autonomously review incoming messages, eliminate spam or irrelevant communication, and evaluate the quality of potential leads by cross-referencing internal data and conducting external research on target companies. Once it has gathered sufficient information, it drafts personalized, contextually relevant responses and correctly routes inquiries to the appropriate support channels. Each of these steps is performed with remarkable speed and consistency, allowing the human overseer to focus on validation rather than execution.
Oversight remains integral to the process. The agent’s daily performance is reviewed by a human manager through Slack, where feedback is provided on tone, phrasing, and decision-making. This continuous loop of feedback enables the system to refine its understanding of Vercel’s communication style and gradually improve the authenticity and nuance of its responses. Since its introduction, this learning framework has enabled Vercel to condense its original ten-person inbound team down to one employee who supervises the agent, while redirecting the remaining nine representatives toward outbound and prospecting roles—tasks that require strategic thinking, communication, and high-touch human judgment.
Reflecting on the broader implications of the project, David Totten, a technology veteran formerly of Databricks and Microsoft and now Vercel’s Vice President of Global Field Engineering, noted that modeling performance after top employees has always been standard corporate practice. What separates the current moment from the past, he observed, is the unprecedented speed and precision with which technology can now capture, analyze, and scale that excellence. Training an AI agent to think and act like a company’s best performer essentially transforms mentorship from a human-to-human process into one that can be digitally replicated across an organization.
Grosser offered a compelling analogy to describe this evolution. The method of teaching AI agents, she said, bears striking similarities to the way companies used to train interns. In both cases, the goal is to observe, emulate, and internalize the behavior of an exemplary worker. No organization would choose to pair an intern—or, by extension, an AI model—with an employee who lacked commitment, failed to understand company values, or exhibited poor performance. Instead, new learners are placed alongside the best—those who embody the culture, demonstrate reliability, and consistently deliver superior results. That same logic now applies to machine learning within the corporate environment.
Importantly, both Grosser and Totten have emphasized that Vercel’s growing reliance on AI is not part of a downsizing initiative. Contrary to fears that automation will lead to widespread layoffs, the company has actually expanded its workforce over the past year. Employees displaced from automated positions have been reassigned to roles that demand advanced human insight, emotional intelligence, and strategic leadership. For Vercel, automation is not about reduction—it is about redirection and enhancement.
Currently, Vercel operates six unique AI agents, all of which are modeled on human best practices observed internally. Over the next six to twelve months, the company aims to scale that number into the hundreds, effectively creating a digital ecosystem of specialized entities, each designed to replicate predictable and clearly defined workflows. Grosser explained that the strongest candidates for automation are processes that are both replicable and deterministic—meaning they produce consistent results when the same inputs are presented. Wherever creativity, ambiguity, or judgment are required, humans will remain indispensable.
Ultimately, this philosophy aligns with Grosser’s personal belief that most people possess untapped capacity for creativity and innovation, but that traditional job structures often constrain these abilities by anchoring them to monotonous, procedural tasks. By offloading such responsibilities onto intelligent agents, she argues, companies like Vercel can empower their employees to explore more intellectually demanding and imaginative pursuits—the kind of work that defines and differentiates human intelligence in an increasingly automated world.
Sourse: https://www.businessinsider.com/ai-agent-entry-level-sales-jobs-vercel-2025-10
