The Automation Paradox: Why AI Could Add More Jobs

Lachlan Reed

It is Monday morning. A mid sized consulting firm announces it has automated nearly every knowledge workflow using AI agents. Investors celebrate. Employees quietly assume layoffs are coming. Six months later, instead of shrinking from two hundred employees to eighty, the company now employs three hundred and forty people. Wait, that is not supposed to happen. Every headline we read tells us AI replaces workers. So why are some of the most AI native companies hiring faster than ever?

Lachlan Reed

Most executives believe that higher productivity automatically means fewer people. They are wrong. It is the story everyone thinks they know. Every new technology creates fear. The loom, the computer, the internet, and now AI. But businesses do not exist simply to reduce costs. They exist to create value. When capability expands, ambition expands. If you build a highway that cuts travel time in half, people do not travel less. They travel more. Businesses expand farther. Entire industries appear that could not exist before. The same thing is happening with AI.

Lachlan Reed

Think about Dan Shippers observation of one of the worlds most AI native companies. They automated almost every workflow. But instead of reducing staff, they expanded from roughly four people to nearly thirty while continuing to hire. What are they hiring those people to actually do? AI dramatically lowers the cost of competent work. Writing, coding, research, marketing, design drafts, and customer support. But lowering the cost of competent work increases demand for exceptional work. Competence becomes abundant. Judgment becomes scarce.

Lachlan Reed

Imagine everyone suddenly owns a professional camera. Photography does not disappear. Instead, people suddenly want better photographers, better editors, better storytellers, and better directors because average quality becomes free. This is the automation paradox. When average execution becomes cheap, the premium on human judgment and exceptional quality skyrockets.

Lachlan Reed

This brings us to a framework we call the Human Sandwich. Every AI workflow has three layers. At the beginning, humans decide what problem matters, what success looks like, and what constraints exist. In the middle, AI performs the execution: searches, drafts, builds, and analyzes. At the end, humans judge, approve, reject, prioritize, and adapt. AI compresses the middle execution, but humans must still define the purpose and evaluate the results.

Lachlan Reed

So why would not companies simply keep the savings? Because businesses rarely stop when they become more productive. Instead, they attempt things they previously could not afford. They launch ten products instead of two. They enter five new countries. They personalize millions of customer experiences. They create new services, expand compliance, and build entirely new business models. It is just like cloud computing. When servers became cheaper, companies did not buy less computing power. They built thousands of new applications.

Lachlan Reed

On top of that, every new AI system introduces entirely new work. AI governance, agent supervision, quality assurance, risk management, policy design, operational resilience, and agent maintenance. Organizations do not eliminate complexity. They relocate it. While some repetitive jobs will absolutely disappear, and some companies will automate purely to reduce cost, the most successful companies automate to increase capability. The question is not whether AI will replace jobs, but what new work appears after the automation.

Lachlan Reed

We have seen this before. Automated teller machines reduced routine bank teller transactions, but banks did not disappear. Many shifted tellers into advisory, relationship, sales, and customer support roles as branch networks expanded. History shows technology changes the composition of work more than eliminating it outright.

Lachlan Reed

So what becomes uniquely human in this new era? It comes down to three things: curiosity, judgment, and leadership. Curiosity, because humans decide which problems deserve solving. Judgment, because AI predicts while humans must choose. And leadership, because organizations still need trust, culture, ethics, vision, and responsibility. None of those can simply be automated.

Lachlan Reed

The company in our opening story did not hire more people despite AI. They hired more people because AI allowed them to imagine work that was not economically possible before. The future is not humans versus AI. It is humans supervising an expanding workforce that just happens to include digital employees. The biggest career risk may not be learning AI. It may be believing AI ends the need for people. Every time humans dramatically increase capability, we invent entirely new work. So perhaps the real question is not whether AI creates unemployment, but whether you are preparing for the kinds of work that do not exist yet.