Change Is How You Keep It

"He who refuses to change will lose even what he seeks to preserve." — Gustav Heinemann

Someone had written it on a whiteboard in a pump-maker's service centre. Not framed, not printed — written in marker, the way you'd note something you didn't want to forget. Heinemann said it as Federal President in 1969. In a workshop that smells of machine oil and decades of accumulated knowing, it lands differently.

The companies I most admire in German manufacturing are not conservative in the pejorative sense. They are conservative in the original one: they conserve things. The master machinist who knows by sound alone when a lathe is cutting wrong. The service engineer who has seen every failure mode this pump can produce, in every condition a customer has ever put it through. The scheduler who keeps production running when three variables break simultaneously, because she has seen it before and knows which lever to pull first. That knowledge is the competitive advantage. It is not written down. It lives in people — and people retire.

This is the problem that AI integration is actually for, when it is done well. Not to replace the machinist. Not to make the service engineer redundant. To ask the right questions before the retirement party, to listen carefully, and to turn forty years of gut-feel into something the next person can learn from. A system that surfaces the pattern. A process that makes the tacit legible.

At Apuna we call this Process Analysis. The work is not glamorous: it is long conversations, careful documentation, iterative testing of whether we have actually captured what matters or only what was easy to capture. The output is not a dashboard. It is the organisation's memory, made searchable and teachable.

What makes this preservation rather than disruption is the design principle behind it: a human always decides. The AI surfaces the pattern; the senior engineer confirms or corrects it. The AI drafts the diagnostic sequence; the service manager approves it before it reaches the field. The tool is a partner and a door-opener — it opens access to knowledge that was previously locked inside one person's head. It is also a limit: the boundary of what is automated is drawn by the people who understand the work, not by the technology.

Heinemann's line lands in a pump-maker's workshop because the people who build and maintain precision equipment understand, intuitively, that reliability is not a property of machines. It is a property of the people who understand the machines, and of the systems that allow that understanding to be passed on. Change the system thoughtfully, and you keep the reliability. Refuse to change it, and the knowledge retires with the person.

The quote on the whiteboard was not a warning about AI. It was written long before this particular technological moment. But it fits the moment exactly: the companies that will preserve what they have built are the ones that find a way to carry their knowledge forward. The ones who refuse — who leave everything in a single master machinist's head because that is how it has always been done — will discover that Heinemann was right.