Guest article: AI doesn't do the work for us, it reveals whether we know our job.

AI and competence are today inseparable concepts in modern knowledge work.

AI has quickly become an obvious part of knowledge work. The tools are faster, more accessible and more powerful than many had anticipated. Production has become frictionless. And that is precisely why one thing is becoming increasingly clear: AI does not replace expertise. It exposes the lack of it.

When AI is used superficially, it acts as a shortcut. When used consistently, it acts as a stress test. It quickly reveals who understands their subject, their context, and their responsibilities, and who is just producing output.

In environments where AI is fully integrated, it is no longer possible to hide behind pace, volume, or pretty form. When everything moves faster, weak strategy, unclear goals, and poor judgment become immediately visible. AI doesn’t fill in the gaps. It magnifies them.

Human work has not become less important. It has become more concentrated. Demarcation, prioritization, perspective and context are now what determine quality. Knowing what not to do is often more important than producing more.

AI is very good at answering questions. It is bad at determining whether the question is asked correctly.

This is where many organizations go wrong. AI is introduced as a production tool before the governance is in place. The result is not laziness but overproduction. More documents, more analyses, more formulations but fewer decisions. The flow increases but the direction is missing.

Used correctly, AI instead functions as a quality requirement. It forces clear goals, explicit assumptions and accountability. Unclear reasoning can no longer be hidden behind work effort. When something is lacking, it quickly becomes clear where the problem lies.

This places new demands on leadership. Not technical demands but demands for intellectual discipline. What has been decided. What is open. What is just a basis. What should actually be done.

The same applies to skills development. AI can be an effective training tool, but only if the requirements remain. The person using AI must be able to explain why an answer is reasonable, see what is missing, and stand by the conclusion. Without that, the tool becomes a substitute for thinking, not a support for it.

In practice, AI acts as a litmus test. It reveals whether an organization has thought clearly before it produces and whether it is prepared to take responsibility when production becomes cheap.

Now that the AI hype has died down, what remains is what has always created value: judgment, prioritization, and responsibility. Technology changes the pace. It does not change the demand for craftsmanship.

_{Of Simon Wallin co-founder Crux Comms}