Blog

For a while now, assembling a screen from an existing design system hasn't taken much time and effort. Consistent components, tokens, patterns, the pieces already fit together before anyone touches them. What used to take craft now often just takes assembly, and increasingly, an AI does that part fine on its own. NN/g's State of UX 2026 report is direct about it: work that's just slotting together existing components is already within reach of the tools doing it faster.

A critique can go great in the room and still slowly wear down the habit of giving feedback, and the reason usually has nothing to do with the session itself.

Some of us remember the time of being just "designers" and moving fast: a sketch, a couple of rounds, a product owner's gut call on what shipped. Then UX became its own discipline, and speed gave way to discovery phases, personas, mood boards, and a rationale deck for every button. The work got more convincing. It also got slow, sometimes to a crawl with projects rescheduled around the time it took to build a persona pretty enough to survive a stakeholder review.

A design recommendation and the reason it gets rejected are often answering two different questions. The recommendation is built around what's right for the user or the product. The rejection is usually about something the pitch never addressed: a budget it doesn't fit, a preference already set before the conversation started, a deadline it doesn't protect, or a decision the stakeholder won't be able to defend later if someone above them asks about it.

You ask the user if the task was simple.

They pause. Mouth slightly open, no sound yet. Eyes drop to the floor for a moment. Small crinkles appear at the temples. Then the smile comes, and the answer: "It was easy."
You know it wasn't.

So, you adopted AI tools. How's that working for you?
Expectations gone up? Headcount justification gone down? The brief is getting shorter because "you have AI now"? Timeline compressed because "AI handles the repetitive work"? And when the output isn't what anyone hoped for - the rushed, under-prompted, under-reviewed work that was due yesterday - it gets written up as a design quality problem?