My pressure cooker has quietly taken control of dinner.

This was not part of a deliberate household transformation strategy. There was no implementation plan, behaviour-change framework or series of stakeholder consultations.

I put ingredients into a machine, press a button and leave.

The pressure cooker assumes that I will become distracted. It expects competing priorities, hungry children, unfinished work and the possibility that I will forget about dinner entirely.

And yet, despite all this predictable human failure, it still produces soup.

This may be one of the most sophisticated systems operating in my life.

Many of the systems we build at work function in precisely the opposite way. They depend on everyone remembering every step, finding the correct document, interpreting ambiguous instructions, contacting the right person and following up at exactly the right time.

When the system fails, we blame the human.

They should have read the guidance.

They should have known who to contact.

They should have followed up.

They should have understood the process.

The pressure cooker makes no such demands. It does not require me to become a different kind of person before I can use it successfully. It adapts to the person standing in the kitchen.

Good systems reduce activation energy. They make the desired action easier to begin and more likely to survive ordinary interruption. They recognise that distraction, uncertainty and competing demands are not exceptional circumstances. They are the conditions under which most people live and work.

A system that functions only when everyone is highly experienced, fully informed, adequately resourced and paying perfect attention is not a robust system.

It is a wish.

It still produces soup.

We often treat exceptional human effort as evidence that a system works. Someone remembers the missing approval. Someone knows the unofficial pathway. Someone stays late to repair the spreadsheet. Someone quietly carries the knowledge that prevents the entire process from collapsing.

The outcome is delivered, so the invisible rescue work disappears.

Then we call the system successful.

The pressure cooker offers another model. The complexity is contained within the infrastructure rather than transferred to the person using it. It does not ask me to monitor the temperature, regulate the pressure or understand the thermodynamics.

It simply lets me make soup.

Perhaps this is the real test of good infrastructure: not whether it can produce an outcome under ideal conditions, but whether it can still produce one while ordinary humans are being ordinary humans.

Good systems do not demand that people become less distracted, less tired or less human.

They begin there.