150 AI Agents, 15 Humans: Working Without Coworkers?

TVS News AI, editorially supervised, presents 150 AI Agents, 15 Humans: Working Without Coworkers.

For Dhruv Amin, a normal working day does not begin with a glance at a full inbox, but with a conversation with dozens of digital colleagues. His company Anything relies on 15 human employees and around 150 AI agents. The ratio is remarkably clear-cut — and, according to Amin, it could soon become more common.

Unlike human colleagues, these helpers are not people, but software. They complete tasks without human intervention. This makes the term “colleague” seem somewhat confusing at first glance: At Anything, one AI handles requests for appointments, another helps with technical problems, and yet another supports customer service.

This gives a familiar workplace question new urgency: Can a company really operate more efficiently when ever more tasks are handed over to autonomously acting software? The Guardian’s report provides an unusually concrete example — and raises an even more important counterquestion: Who benefits from this change?

150 Agents Share the Working Day of 15 Humans

The Staffing

Anything’s numbers are the report’s most striking finding. There are approximately 150 AI agents for every 15 human employees. Amin believes such a ratio could become the norm in the future. In doing so, he is describing precisely the future that several start-ups are pursuing: They are developing agents that are not merely meant to answer individual questions, but to carry out entire workflows independently.

At Anything, the areas of responsibility range from routine tasks and support to activities that would otherwise clearly be assigned to human colleagues. These include office organization, technical support, online marketing, customer service and even restocking the kitchen.

The Software Works Independently

The crucial difference lies not in an agent’s name, but in its function. The systems are software that completes tasks without human intervention. A human enters an instruction or message; the appropriate agent then handles matters in its area. This makes the model more than just a chatbot that formulates advice, turning it into a distribution of tasks across several specialized digital helpers.

A Working Day with Gwen, Cody, Betsy, Juno and Justin

Appointments and Technical Questions

On an average working day, Amin chats with dozens of his AI colleagues. When he needs an appointment, he turns to Gwen, his assistant for executive support. If he encounters a technical problem, he asks Cody, his programming partner. Their names and roles create an almost human workplace atmosphere, even though software stands behind every conversation.

Marketing and Customers

Betsy is responsible for search engine optimization questions. Juno supports Amin when onboarding new customers. The digital colleagues therefore do more than cover administrative tasks. They intervene in typical activities involving sales, marketing, technology and customer contact.

Even the Kitchen Runs Digitally

Justin is the digital office manager for the kitchen. If snacks are running low, Amin sends him a Slack message. This final example illustrates just how far the division of labor extends at Anything: The boundary between the physical workplace and digital assistance becomes blurred.

More Output with Fewer People — and Where the Conflict Begins

The Start-ups’ Promises

Anything is one of several start-ups developing AI agents. Their shared goal is a working world in which ever more tasks in sales, marketing, programming and other fields are handed over to semi-autonomous AI systems. The benefit is supposed to be greater output with fewer staff.

This is precisely where The Guardian sees a tension. Companies advertise the ability to accomplish more with fewer resources. At the same time, pressure is building on human employees whose existing tasks, responsibilities or workload are redistributed. Amin does not describe his company as a battlefield, but the ratio of 15 people to 150 agents illustrates how far the change has already reached into a very small business.

The precise question, therefore, is not whether AI can fundamentally replace people. The report presents a concrete working day in which people assign, delegate and receive results from software. What remains unclear is how those involved experience their roles under this system — and who receives the gains from greater efficiency.

What the Work Model Means for You in Practice

More Tasks, Less Routine Work

For employees at a company like Anything, the change initially means a new division of labor. Appointments, technical support, search engine optimization, customer onboarding and keeping the kitchen stocked are handled by specialized agents. People are therefore no longer personally responsible for every task, but instead manage a broad system of digital helpers.

This can make work easier by allowing routine tasks to be handed over to software. But The Guardian also makes the other side visible: When companies pursue the use of agents with the goal of achieving more with fewer staff, that directly affects the working world of the employees who remain.

For anyone interested in such a model, Anything therefore offers an extraordinary insight into the working world of tomorrow. Not an abstract vision of the future, but a company where 15 people already organize a normal working day with 150 software agents. The Guardian thus provides a concrete starting point for the most exciting question: Who wins when artificial colleagues take over tasks — and who bears the consequences of efficiency-driven pressure?