Your Process Got Compressed. Your Thinking Shouldn't.

There's a seductive idea gaining traction: AI means you can skip the messy middle of design. Why research when you can generate? Why ideate when the machine iterates? A new piece from Nielsen Norman Group pushes back hard on this narrative — and a pair of typography stories quietly proves why it's right.

What Happened Today

The process isn't dead. It's denser.

NNGroup's latest article [NNGroup] tackles a misconception that's been building for months: that AI makes the design process obsolete. Their argument is more nuanced and, frankly, more useful. The process isn't disappearing — it's being compressed.

Here's the key insight: when experienced designers appear to skip steps, they haven't actually eliminated anything. They've internalized years of research, pattern recognition, and user understanding so thoroughly that their decision-making looks like intuition. It's not. It's deeply informed judgment operating at speed.

AI can accelerate each phase of design — research, ideation, prototyping, testing — but only when guided by someone who understands what each phase is supposed to accomplish. Strip out that understanding and you're not compressing anything. You're just cutting corners faster.

Craft as compressed knowledge

Two typography stories this week illustrate what internalized process actually looks like in practice.

Studio Drama's new custom typeface for Snickers [Creative Bloq] captures the brand's irreverent personality through deliberately irregular, "nutty" letterforms that echo the chunky texture of the product itself. It's playful, it's specific, and it's the result of deep brand thinking expressed through typographic craft. An AI could generate a hundred display fonts in the time it took to develop this one. None of them would carry this level of intentionality.

Meanwhile, typographer Josse Pickard [Creative Bloq] reveals an inspiration palette spanning circuit boards, weathered stone, and brick lettering — pulling from architecture, electronics, and material decay to inform his letterforms. This kind of cross-disciplinary synthesis is exactly what NNGroup means by internalized process. Pickard isn't following a checklist. He's drawing on a deep reservoir of observation and material knowledge that makes his work feel both technically precise and organically alive.

Neither of these projects could be meaningfully accelerated by AI, because the value is the judgment.

Persuasive design: what actually survives compression

Anders Toxboe's retrospective on persuasive design [Smashing Magazine] adds another dimension to the compression conversation. Revisiting the field after a decade, Toxboe finds that many product teams still over-rely on surface-level behavioral nudges and superficial gamification — the design equivalent of skipping to the answer without understanding the question.

The patterns that have held up over ten years are rooted in genuine understanding of human motivation, not quick tricks. Toxboe's distinction between lasting principles and short-lived hacks maps directly onto the compression argument: you can speed up how quickly you apply a persuasive pattern, but you cannot shortcut the understanding of why it works and when it's appropriate.

Tool Spotlight: CSS corner-shape

On the implementation side, the new CSS corner-shape property [Smashing Magazine] is a small but satisfying example of good compression in action. For years, achieving beveled, scooped, or squircle corners meant brittle workarounds — clip-path, SVG masks, and other fragile hacks that broke the moment you looked at them sideways. Now the browser handles that complexity natively, letting designers and developers express intent directly.

This is what healthy compression looks like: the underlying complexity doesn't vanish — it gets absorbed by better tooling so practitioners can focus on what they want rather than how to hack it together. It's the CSS equivalent of NNGroup's argument. The process is still there; it's just been internalized by the platform.

The Bigger Picture

The thread connecting today's stories is the difference between compression and elimination.

AI compresses timelines. Better CSS compresses implementation. Experienced practitioners compress decision-making. But in every case, the underlying knowledge, rigor, and judgment are still doing the work. They're just less visible.

This matters because the current discourse often conflates speed with simplification. "AI can generate a wireframe in 30 seconds" gets treated as proof that the thinking behind wireframing was trivial all along. It wasn't. The wireframe was never the hard part. Understanding what should be in it was.

I think this is where a lot of the anxiety in the profession is coming from. When AI makes outputs cheap, people mistake that for the work itself becoming cheap. But the work was never the output. The work was the thousand micro-decisions that shaped it.

Takeaway

If your design process was shallow before AI, AI just makes it shallower faster. If it was deep, AI genuinely makes you more powerful.

The question worth asking yourself this week isn't "How do I go faster?" It's: "What have I internalized deeply enough to compress — and what am I just skipping?"

Compression rewards depth. Invest accordingly.