An AI-Native Company Can't Run on an Old Org Chart

If AI lets one person do work that used to take five, why are most companies still organised as if nothing has changed?

That's the question I want to ask you, and one I have spent a lot of time reflecting on.

Over the last couple of years, AI has become part of how I think, write, learn and solve problems. The more I used it, the more capable I became. I could learn faster, draft quicker, pressure-test ideas and cut hours off work that used to take all day.

I encouraged my leadership team to work the same way.

And this exploration led us to a different question: if individuals are becoming much more capable, shouldn't the structure of the organisation change too?

Hierarchy was one of humanity's great inventions. It let organisations scale when information was scarce, coordination was expensive, and context lived unevenly across a company. Managers connected the dots. Titles clarified authority. Decisions moved through layers because they had to.

It was the right model for the world we were operating in. I no longer think it's the right model for the world we're entering.

Today, one person can access information that used to live across multiple functions, analyse data, build prototypes, write code, draft strategy and test ideas, all in hours instead of weeks. So why should work still move through the same layers it always has? If the person closest to a customer already has the context and the tools to solve a problem, why make them wait for three approvals first?

These questions led us to what we now call Flatland.

People assume Flatland is about removing titles. Honestly, that part was easy. Flatland is really about redesigning how an organisation works.

For decades, organisations existed to move information and decisions up and down layers. AI changes that. When context and intelligence are available to nearly everyone, an organisation's job shifts: less about distributing information, more about helping good people solve hard problems together, without needing permission first.

Here's what that looks like in practice. A Builder at one of our hubs notices customers keep getting stuck at the same step. A few years ago, they'd have written it up, sent it to a manager and waited for it to work through a few rounds of review, hoping someone eventually acted on it. Today, that same Builder can investigate the problem, prototype a fix and bring it back to the team within hours.

That's the organisation we're trying to build, which is also why everyone at Cars24 is simply called a Builder. We wanted contribution to be the only signal that mattered here.

Though, none of this came from simply redrawing the org chart. Companies build habits around hierarchy. People get attached to titles, reporting lines, approval chains, and the comfort of knowing exactly where a decision comes from and where it goes next.

We spent months debating what Flatland would actually take. How do you give people more autonomy without losing accountability? How do you recognise real contribution without falling back on hierarchy to do it for you?

Those conversations mattered because they forced us to think beyond organisation charts. They gave us clarity on the behaviours we wanted to encourage and the ones we needed to leave behind.

That's why we spent just as much time changing how people worked as we did changing the structure itself. Every Builder was expected to make AI part of their daily work, build workflows, automate the repetitive stuff, experiment constantly, and cut out dependencies wherever possible. If one person could now do what used to take five, we didn't want five people still coordinating around it out of habit.

The same expectation applied to leadership. Companies don't become AI-first because the founder talks about AI a lot. They become AI-first when leaders start working differently: building more themselves, delegating less, and clearing friction instead of creating it. Culture follows what people do, not what gets announced.

These thousands of small behavioural changes happened first over months. People became more self-sufficient, leaders became more hands-on and decisions started moving differently. Only then were we able to call ourselves Flatland and tell the world about it.

Lastly, I don't think changing an organisation is a one-time exercise. You don't redesign how people work and call it done. You keep learning, adjusting and refining what's working and what isn't.

Every generation gets to question one assumption the last one took for granted. Ours might be this. We don't have the final answer. But we'd rather build for the world that's coming than optimise for the one that's disappearing.