Stop working harder, work smarter.

There is a realisation I arrived at too late. Not because it is difficult — because it is uncomfortable.

I am not a bad delegator. I am a founder who did too much himself for too long. Not out of a need for control — out of poor system design. I had no clean definition of which work is *my* work and which is not. So I did everything. And everything costs time. And time is the one thing a solo founder cannot buy.

Eventually the bill arrives.

The solution was not a question of effort. It was a question of order.

**Eliminate → Simplify → Automate → AI.**

In that order. Not as a slogan — as a logic for decisions.

Anyone who starts directly at AI has already lost. Not because AI is bad. Because they have just automated the wrong thing — immediately, with leverage, and at a scale that was not previously possible. If the process is wrong, an AI tool makes it wrong faster. That is not optimisation. That is scaling errors.

So I start with eliminating.

What happens if you remove something? Often: nothing essential. Many processes exist because someone introduced them at some point and no one ever asked whether they are still needed. The honest answer is frequently: no.

What remains gets simplified. Not for the sake of simplicity — because complexity creates friction, and friction costs energy that is needed elsewhere.

What remains after that, and is deterministic — meaning: clearly defined input, clearly defined output, no real judgment needed — gets automated.

And what is complex, context-dependent, not fully specifiable: that is where AI arrives. Not before.

The second realisation was the more uncomfortable one.

My time costs money. Not in an abstract sense. In the concrete sense: if I do something I could delegate, I pay a real opportunity cost. I gave myself a rule — which I will write here the way I wrote it to myself, without false modesty and without arrogance:

*I only do work worth ≥ €200 per hour to a client.*

Everything below that goes to agents.

That sounds harsh. But it is not a question of pride — it is a question of sustainability. I work with finite cognitive capacity. Attention is not infinitely renewable. If I spend hours answering emails an agent would handle better, formatting reports a script would format more cleanly, or gathering data points a tool would find faster — I am spending capacity that is needed for the work that actually requires judgment. Client conversations. Strategic decisions. The assessment of whether a technical concept genuinely works for a specific organisation.

The €200 threshold is not arbitrary. It is the point at which my work stops being a production input and starts being an investment.

Taoism has a principle I understood late as a system design principle: Wu Wei — action without resistance. Not inaction. Not passivity. Finding the path of least resistance — the path on which the right thing happens with the least force.

That sounds like philosophy. It is actually a design task.

A well-built system does the right thing almost by itself. A poorly built system fights its own structure constantly. Anyone who notices they are doing the same work by hand every day, rebuilding the same decision from scratch, explaining the same problem for the third time — that person is fighting the structure. Not external obstacles. Their own system design.

Wu Wei is the signal: the problem is not insufficient effort. It is the design.

Delegating to agents is not magic. It requires design work.

An agent that is poorly briefed produces poor results quickly. An agent that is not calibrated makes wrong decisions over time. And an agent operating unsupervised in a high-stakes area is not a lever — it is a risk factor.

The €200 rule is therefore also a responsibility rule.

I delegate the work that requires no judgment. But I retain the work where judgment is decisive — and with it, the responsibility for it. No agent makes the assessment of whether a recommendation is right for this client at this moment. No agent signs for the consequences. That is my work. And I do it — because I have been freed up for it.

That is the difference between automation as escape from responsibility and automation as the condition under which responsibility can actually be carried.

Apuna is early. No large client portfolio, no three-sentence success story I could trim for this post. What I have is the operating system itself — the practice I have described here.

What I can give clients is the same operating system for their work.

Not the software. Not a toolbox. The principle: eliminate first. Then simplify. Then automate. Then AI — with the human at the point where judgment is needed.

That is what Apuna does. Not as a promise. As a method.