UX Collective: Craft still matters, but it's about outcomes

Patrick Neeman, AI builder and Senior Principal UX Designer, published this piece in UX Collective in July 2026 to challenge a persistent assumption: that design craft still lives in production quality. His argument is that AI has already automated production fluency, so the locus of craft has shifted — and designers who haven’t noticed are measuring themselves against the wrong standard.

The previous decade’s benchmark was artifacts: pixel-perfect Figma files, tight design systems, handoff-ready components. Neeman’s provocation is blunt: “A Figma frame has never shipped revenue.” The artifact was always a proxy for something else, and AI now produces proxies faster than any human team.

The more dangerous version of the same problem shows up in research. AI tools can generate personas, synthesis readouts, and usability test summaries that look credible but are invented. The document finishes before actual research happens. Teams then act on it. Neeman calls for “a research habit that feeds the machine” — structuring real inquiry so that AI tools have something genuine to accelerate rather than patterns to hallucinate.

What he identifies as craft in the AI era is quieter and harder to display: the judgment to know when a generated output is wrong, the skill to describe a standard clearly enough that an AI can follow it, and the discipline to keep humans reviewing outcomes rather than just approving outputs.

The article ends with a counter-intuitive claim: the same AI tools that threaten shallow craft can also protect genuine craft if you know what you’re protecting. That requires having defined standards before delegating production to the machine.

This is worth reading for mid-to-senior designers who feel that AI has devalued their skills but haven’t named what’s been devalued or what remains. Neeman provides a frame for that distinction, along with concrete habits that make the distinction matter in practice.