No HR, no employees: Ex-eBay worker builds AI company with 27 agents, no human staff

A former eBay employee who lost her job after 11 years at the company has now built an AI-powered marketing business that operates without traditional employees, HR teams, or managers. In an interview with Business Insider, the founder revealed that her company now runs using 27 custom AI agents that handle almost every part of the business, including market research, ad campaigns, analytics, creative work, and customer conversions.

She said the idea came during a difficult phase in her life. Shortly after moving with her family from Switzerland to the US, she was laid off from eBay at a time when the tech industry was seeing widespread job cuts.

Instead of re-entering a crowded job market, she decided to experiment with building something of her own. A few months later in 2024, she launched a small marketing agency, despite not having a formal marketing background.

Her experience in analytics at eBay, however, helped her understand how to structure AI workflows and automate repetitive processes. As tools like OpenAI’s ChatGPT became more advanced, she began exploring autonomous AI systems where multiple agents could work together with minimal human involvement.

According to her, the company now functions through a three-layer AI structure. One layer defines how the agents operate, another decides which AI handles specific tasks, while the execution layer performs the actual work.

At the centre of the system are six core AI agents focused on areas such as research, finance, legal review, data analysis, creative direction, and workflow coordination. Once tasks are assigned, other AI agents handle technical work, traffic generation, ad optimisation, and revenue-focused activities. She also revealed that the entire setup costs less than $1,000 a month to maintain, including subscriptions for Claude Code, ChatGPT, APIs, and AI tools such as HeyGen and ElevenLabs.

Before onboarding paying clients, she reportedly tested the workflow using 14 client profiles, including personal projects and free work for friends, to make sure the system could consistently deliver quality results. Today, the AI agents reportedly run advertising campaigns, analyse performance, improve creatives, and generate reports automatically. Once a client is onboarded, she spends only around two hours a week overseeing the account.

Microsoft's warning about AI replacing humans jobs already coming true

The story has also reignited concerns around AI replacing white-collar jobs, especially after recent comments from Mustafa Suleyman, who leads AI efforts at Microsoft. In an interview with the Financial Times, Suleyman warned that AI could automate “most, if not all” white-collar jobs within the next 12 to 18 months. He also said AI systems are approaching “human-level performance on most professional tasks.”

While such predictions have often sounded futuristic, examples like this AI-run marketing business are beginning to show how quickly workplace structures could change. Tasks that once required teams of marketers, analysts, copywriters, and campaign managers are now being handled largely by software agents supervised by a single person. Even so, the founder said human expertise still matters. According to her, AI can analyse information and automate workflows, but it still struggles with judgment, emotional understanding, and reading the mood of a client conversation.

She added that domain knowledge remains critical when building AI systems. Without expertise, she said, it would be difficult to create reliable AI tools for industries such as healthcare or law. The founder believes repetitive office work will increasingly be handled by AI in the future, while humans shift into oversight and strategy-focused roles.

At the moment, she still works with freelancers for a few older clients. But every new client is now onboarded through the AI-driven system. She also believes that if she focused only on supervision, she could comfortably manage up to 25 clients on her own, something that would traditionally require a much larger team.