The 1/100th company: AI and shrinking workforce

Can a company that employs ten thousand people today operate with just a hundred tomorrow? In several industries, this is no longer a speculative question.

Can a company that employs ten thousand people today operate with just a hundred tomorrow? Not across every sector, and not immediately. But in several industries, this is no longer a speculative question. It is an emerging trajectory.

A recent personal experience brought this into sharper focus. Working with Claude on coding and reasoning tasks, I found myself pausing more than once. The system was not merely assisting. It was structuring problems, proposing approaches, and iterating with a level of coherence that suggested something more than incremental progress. It felt like a shift to a higher gear in the AI journey.

If such capabilities continue to evolve, the implications for organisations are significant. Firms have historically scaled up to meet the growing demand for peoples’ skills like coding. Also, bringing people together, aligning them, managing workflows, and ensuring accountability required layers of management. That logic may be changing. As coding becomes automated and coordination becomes computational, the need for large numbers of people begins to diminish.

This does not mean that organisations will disappear. It does mean that many of them may shrink. In information-heavy sectors such as software, media, analytics, and parts of finance, a small team augmented by AI systems could deliver what once required entire departments. The “1/100th company” is not a fantasy. It is a design possibility.

The timeline is not immediate. In physical sectors such as healthcare and infrastructure, human presence and safety accountability will remain essential. But over the next decade, even these sectors will experience a compression of workforce in areas where decision-making, monitoring, and optimisation can be automated.

This shift will not be limited to organisations. It will reshape work itself. The familiar structure of the eight-hour workday and five-day work week is unlikely to hold in its current form. It was built for a world where output scaled with the time spent. In an AI-augmented environment, output scales with leverage. A few hours of well-directed work may achieve what previously required days. The constraint will no longer be time. It will be clarity of thought and quality of inputs.

This brings us to a critical shift in value. Domain experience will become central. As AI systems grow more capable, the limiting factor will not be their ability to generate responses, but the quality of questions that are posed to them. What is often referred to as prompt engineering is, in essence, the articulation of context. That context comes from experience. Those who understand the nuances of a domain will be able to direct AI far more effectively than those who do not.

Careers, therefore, may take on a different shape. Early years may involve more structured roles where individuals build a foundational understanding. But as experience accumulates, work may fragment into projects and engagements. The trajectory could shift from stable employment to a portfolio of gigs, each drawing on accumulated expertise.

The most successful professionals in such a world will likely share a few characteristics. They will stay close to sources of information. They will engage directly with customers. They will remain hands-on, resisting the temptation to move too quickly into abstraction. In a system where much of the execution is automated, proximity to reality becomes a source of advantage.

There is, however, a paradox at the heart of this transition. If AI systems can perform a large share of the work, how do individuals gain the experience required to guide those very systems? Experience is not acquired through observation alone. It is built through doing, through mistakes, through context that cannot always be codified.

Yet, these very experiences are what feed AI systems, both directly through data and indirectly through human guidance. This creates a feedback loop. New forms of experience will need to be designed. Apprenticeship, simulation, and hybrid human-AI workflows may become critical in ensuring that expertise continues to develop.

The 1/100th company is about a different balance between people and systems. It is about organisations that are less constrained by coordination and more defined by purpose, judgment, and accountability. Some companies will shrink dramatically. Others will not. But across sectors, the assumption that scale requires proportional headcount is beginning to weaken.

The more important shift lies beneath the numbers. A company may increasingly be defined by how effectively it combines human insight with computational capability to deliver outcomes.

Disclaimer: The views expressed above are the author's own. They do not necessarily reflect the views of DH.