How Designers Cultivate the Soil for Good Judgment

I recently came across a design decision in an NN/G article on Accelerating Research with AI that felt remarkably human.

Facing the same "transaction failed" error screen,

AI wrote: "Transaction failed, please try again."

The designer wrote: "We couldn't complete this transaction, but your funds are safe—no charges were made."

The difference between these two sentences isn't a matter of UI technique or copywriting skill. It's an understanding of what the user is afraid of in that exact moment.AI couldn't write the second sentence—not because its language ability is poor, but because it has no idea what prior experience the user is carrying, or what they're afraid of in their heart, the moment they hit submit. 👻

🌱 The soil of judgment: contextual understanding

So here's the question—where does a designer's judgment actually come from?

I used to think it was "experience" or "intuition," but I've gradually come to realize that underneath it all is reallycontextual understanding. In other words, how well you grasp why something matters, to whom it matters, and under what circumstances it matters. 🤯

When contextual understanding is lacking, a designer's judgment easily turns into bias—just applying whatever solution they're used to, rather than genuinely responding to the client's problem.

🔍 Research is the mechanism that fills in context

So how do you build contextual understanding? My current answer is:do research.

Research isn't some fancy word—its meaning is actually quite plain: it means going and understanding something. Understanding this industry, this user, this client's business logic. 📖

AI is genuinely helpful during the execution phase of research—it can help organize interview transcripts, summarize key terms, and speed up data processing. I use all of this now, and it really has boosted efficiency.

But an NN/g (Nielsen Norman Group) article put it very directly: don't let AI do the analysis for you. Because AI is stochastic—it may latch onto certain aspects, itwill keep circling back to whatever it happened to notice, whileat the same time overlooking others. It might misread things, or even "manufacture" insights that don't actually exist. To put it more bluntly—if you outsource all of your analysis to AI, what you're gambling with isn't just the quality of the insight, but your own professional credibility.

Analysis and interpretation still need to be done by the designer. Because that's exactly where contextual judgment happens.

🧭 Contextual judgment is like telling a story

A designer's understanding of user needs has to be grounded inspecific context, not abstract, generic assumptions. 🎭

The same feature can mean something completely different depending on the context:

• A user using an app "during their commute" versus "sitting at a desk" are two entirely different design problems
• A "save" button carries a different sense of urgency for someone in a rush versus someone calmly organizing files

So judgment can't exist independently of context.

A good designer doesn't ask "what does the user want," but instead asks "what does this particular user want, right now, in this moment."

✍️There's no shortcut to cultivating judgment