The craft test: Questions that reveal whether you're getting real design or AI commodity
Francois Brill
Founding Designer
There's a famous principle in design: you're done not when there's nothing left to add, but when there's nothing left to take away.
It sounds elegant. In practice, it's brutally hard.
Because adding is easy. Anyone can layer on features, pile up options, accumulate capabilities. AI makes this faster than ever—generate screens, produce variations, ship more.
Subtracting is hard. It requires understanding why each element exists. What serves the user and what creates friction. What reinforces the brand and what dilutes it. What must stay and what needs to go.
That understanding is craft. And as commodity design floods the market, knowing how to spot craft becomes a critical skill for anyone building products.
The flat white principle
I love coffee. Spent real money on a proper home setup because for me, there's a science to it.
Right beans. Right grind for the day's humidity. Exactly 18 grams, tamped with precisely the right force. Extracted at 7 bar over 25 seconds. Milk frothed to the perfect texture. Combined into a flat white.
Most people would enjoy that drink. Very few could tell you how it was made. Which steps mattered. What could be skipped.
They just judge the result. They love it or they don't.
Products work the same way. Users experience the outcome—does it solve their problem or create friction?—completely oblivious to the process that created it.
This is why evaluating craft is hard. The 90% that matters most is invisible. But if you know the questions to ask, you can see beneath the surface.
“Users experience the outcome—completely oblivious to the process that created it. This is why evaluating craft is hard.
Five questions that reveal craft
Whether you're evaluating a design partner, a new hire, or your own team's work, these questions separate craft from commodity:
1. "Walk me through why."
Pick any decision—a layout choice, a feature approach, a visual direction. Ask "why".
Craft answers connect decisions to user needs, business constraints, strategic positioning. There's a chain of reasoning. Each choice ladders up to something bigger.
Commodity answers are vague. "It looked good." "The AI generated this." "We liked this option." No connection to strategy or insight.
2. "What did you remove?"
Real design is subtractive. Features get killed. Flows get simplified. Options get consolidated.
Craft has a graveyard—ideas that were explored and deliberately rejected. Ask about what didn't make it and why.
Commodity adds without removing. If nothing was cut, nothing was truly designed.
3. "How does this connect to the rest?"
Pick an element. Ask how it relates to the broader product, the brand, the user journey.
Craft thinks in systems. Each piece reinforces others. There's coherence across touchpoints.
Commodity treats pieces in isolation. Elements were built separately, assembled without integration.
4. "What would happen if we changed X?"
Propose a modification. See if they can trace the implications.
Craft understands connections. Change this, and that breaks. Modify here, and it affects there. The team has a mental model of the whole.
Commodity can't answer confidently. They assembled outputs without understanding how pieces relate.
5. "Show me the process."
Ask to see the 90% below the iceberg. Discovery findings. Strategic decisions. Options explored. Iterations.
Craft has artifacts. Research summaries. Decision logs. Rejected directions with rationale.
Commodity has deliverables with no backstory. The process was prompting until something looked right.
“If nothing was cut, nothing was truly designed.
Red flags in practice
Beyond questions, watch for these signals:
No single-threaded strategy. As you move through features, there's no clear logic. Each element makes sense alone; together, they don't add up. The product feels accumulated, not designed.
Haphazard feature layering. Capability piled on capability. More options, more complexity, more friction. No one made hard decisions about focus.
Inconsistent patterns. Similar actions work differently in different places. Visual language drifts across screens. You can feel the lack of underlying structure.
Generic experience. Remove the logo and it could be anyone's product. Design follows trends without interpretation. Nothing distinctive about the experience.
Defensive responses. When you ask why, the team gets uncomfortable or deflects. Craft welcomes scrutiny because there are real reasons behind decisions.
Green flags that signal craft
And here's what to look for on the positive side:
Coherent backstory. Every decision connects to something—user insight, business goal, strategic constraint. The team can explain the reasoning and it makes sense.
Visible restraint. The product does fewer things well rather than many things poorly. Someone made the hard calls about what to cut.
Systems thinking. Components relate to each other. Patterns are consistent. You can feel the underlying structure even before it's explained.
Distinctive experience. Cover the logo and you'd still know whose product this is. The design expresses something specific about the brand.
Process artifacts. There's evidence of thinking—research, explorations, rejected directions. The deliverable emerged from genuine process.
“Craft is hard to produce. That's exactly why it's valuable.
Why this matters now
AI has made professional-looking design accessible to anyone. Commodity is easier to produce than ever before.
This makes evaluation harder—and more important.
Users will feel the difference between craft and commodity. They won't articulate it. They'll just love the product or abandon it. Like a perfect flat white, they know quality when they experience it.
Your job is to recognize craft before users render their verdict.
Ask the questions. Look for the signals. The 90% you can't see is what determines whether your product wins or becomes another forgettable entry in a crowded market.
Craft is hard to produce. That's exactly why it's valuable.
Learn to spot it.
