A study published this week showed that engineering teams using AI-enabled workflows can now test nearly 4x more design options per project than conventional approaches.
It's being framed as a win for design velocity. I think it's a trap.
Here's why.
Variants Without Clarity Are Just Noise
Testing more options is only valuable if you know what you're optimising for. Most product teams don't. Not really.
They have backlogs of competing ideas, stakeholder preferences dressed up as user needs, and success metrics that were agreed in a workshop 18 months ago and never revisited.
Give those teams 4x the variant capacity and you don't get better decisions. You get better-looking noise. Design reviews where someone picks the prettiest option and calls it validated. Shipping decisions made by whoever's most confident in the room.
I've seen this play out. At Kraken, working on the consumer app, the constraint was never our ability to produce options. We had people, tools, and time for iteration. The actual bottleneck was always clarity: what does a good outcome look like for this user in this moment?
When that question is answered well, iteration gets fast and purposeful. When it's not, you can run 50 variants and still ship the wrong thing.
AI Lowers Production Cost. Production Was Never the Hard Part.
What AI actually changes is production cost. It lowers the time and effort required to generate design options, copy variants, layout iterations. That's genuinely useful.
But production was never the hardest part of product design.
The hardest parts are: - Framing the right problem in the first place - Knowing when your assumptions are wrong - Deciding which signal matters when everything looks plausible - Pushing back on a brief that will lead the team off a cliff
None of that gets easier with more variants. If anything, it gets harder. Because now you have more options to hide behind. More surface area for the loudest voice in the room to win. More ways to look busy without getting to the real question.
Judgment Beats Speed
Here's what I think separates designers who consistently ship great work from designers who produce a lot of work.
It's not speed. It's judgment.
Judgment is knowing that the onboarding screen isn't actually the problem. Judgment is recognising that the team is solving for the metric, not the user. Judgment is looking at a clean Figma screen and knowing it will confuse someone on day one because you've been in that situation before.
That skill comes from shipping things, watching them fail, seeing users struggle, and recalibrating. It doesn't come from generating more options. It comes from being forced to commit to one and living with the consequences.
AI tools can generate 4x the variants. They can't give you 10 years of watching what happens after launch.
Use AI to Get Clearer, Not Just Faster
So if you're leading a design team, the question to ask isn't "how do we use AI to go faster?"
It's "how do we use AI to get clearer?"
Use it to synthesise research faster. Use it to stress-test a brief before anyone starts designing. Use it to surface edge cases you hadn't considered. Use it to cut the time between "we have a question" and "we have enough to make a call."
But protect the parts of the process that build judgment. Don't automate away critique. Don't skip the awkward conversation where you challenge the premise. Don't replace the designer's ability to synthesise feedback with a tool that just averages it.
The teams that come out ahead won't be the ones with 4x the variants.
They'll be the ones who asked better questions before generating any.
That's the bet worth making.
Fact Check
Every factual claim in this article, with its source.
Claim: AI-enabled engineering teams can test nearly 4x more design options per project than conventional approaches. Source: GlobeNewswire (March 2026), AI-enabled engineering and design workflow study. globenewswire.com
Claim: 91% of designers say AI tools improve their designs and 89% report working faster with AI tools in 2026. Source: Figma, State of the Designer 2026. figma.com
Claim: When leaders prioritise design excellence, designers are twice as likely to feel good about their work. Source: Figma, State of the Designer 2026. figma.com
Unsourced statements (Jay's opinion or lived experience): The Kraken consumer app reference and the framing of clarity as the actual bottleneck; the four-section argument about production cost vs judgment; the practical advice on how to use AI to get clearer, not just faster. These are Jay's points of view, not third-party data.