← Back to the notebook

AI & work

AI makes judgment more valuable.

When execution gets cheaper, deciding what deserves to exist becomes the real work.

AI has made it easier to produce almost anything. Another campaign. Another dashboard. Another product idea turned into a working prototype. The cost of making the first version keeps falling.

That changes where the difficult work sits. Before you can ask a system to build something, you need to decide whether it deserves to be built. You need to know what a good result looks like. And you need enough understanding to recognize when a convincing answer is wrong.

The decision comes first

In marketing, this shows up quickly. A team can generate dozens of versions of a message without resolving who needs to hear it or why they should care. The output looks productive. The uncertainty underneath it stays exactly where it was.

My starting point is the decision. What are we trying to change? What evidence would tell us it worked? Who needs to act on the result? Those questions shape the system before a tool enters the conversation.

Give the system a standard

Delegating execution requires a clear standard. If the brief says “make this better,” the system has to invent what better means. If the brief names the audience, the constraint, the intended action, and the evidence it can use, the work becomes possible to evaluate.

This is why taste and judgment matter more as the tools improve. Someone has to choose the direction, identify the weak assumption, and reject the polished answer that misses the point. Generating more options does not make those decisions disappear.

Build the review into the work

A useful workflow has a place for human review before its consequences become expensive. A factual claim needs a source. A customer-facing message needs a voice check. A system change needs a test of what happens when the inputs are missing or wrong.

That review should be specific. “Does this sound good?” is a different question from “Would this give the buyer an accurate reason to act?” The second question makes the work better.

The opportunity is to spend less time producing the first pass and more time making decisions worth executing. That is the kind of AI-enabled work I want to build.

Thinking is better with a conversation.

Continue on LinkedIn ↗