Happy Wednesday.
We just published the State of Prototyping: Spring 2026 Report - our first open survey on how designers and builders actually work right now.
1,478 responses across 18 regions. What tools people use every week. How much they're vibe coding. Whether they trust AI to ship. And where the real anxiety is.
The full interactive report, the raw dataset, and the API are all open.
Here's what stood out.
– Tommy (@designertom)
MADE POSSIBLE BY
This survey was independently run by UX Tools. Distribution and production was supported by our Spring 2026 sponsors.
Framer. Ship real sites without writing code. Full design freedom, built-in CMS, and the performance your clients actually need.
MagicPath. Describe what you want, get clickable prototypes on an infinite canvas. Design and code in one place.
Dscout. Run user research from recruit to insight - video diaries, interviews, and surveys in one AI-powered platform.
MagicPatterns. AI prototyping that matches your existing product. Import your design system, generate on-brand UI, export to Figma or code.
Mobbin. Stop screenshotting apps. Browse the largest curated library of real product UI - search by flow, pattern, or screen.
The most-used design tool after Figma is now AI
Five of the ten most-used weekly tools are AI. Claude is at 50.8%, which makes it the #2 weekly tool in design after Figma. Claude Code sits at #4, ahead of FigJam.
An AI coding terminal is now more embedded in designer workflows than any canvas-first tool. That happened quietly.
The profession split into thirds.
43.8% of designers spend more than half their time vibe coding. 31.1% say it's most or all of how they build. And 37.7% do zero.
Not "a little." Zero.
These aren't different generations. They work in the same orgs, on the same products, in the same Slack channels. The gap between design engineers at 80.9% and IC designers at 35.0% is the widest split in the data.
More than half of designers are building for themselves.
59.1% have built their own tool, app, or utility with AI in the last six months. One in four does it regularly. Only 10.4% have no plans to.
Two years ago that number was in the low single digits. The ability to build something for yourself — not a product, just a thing that solves your own problem — unlocked something.
The biggest blocker isn't AI quality. It's time.
Time to learn tools (55.7%), too many tools to evaluate (53.0%), and AI output quality (52.2%) are within 3.5 points of each other. The industry shipped a hundred tools and forgot to give anyone time to learn them.
Design engineers feel more valuable. Researchers feel most at risk.
Design engineers: 50.0% feel more valuable, 10.6% feel less secure.
Researchers: 17.4% feel more valuable, 39.1% feel less secure.
Same industry. Opposite experience. If your org isn't talking about this openly, it's not because the problem isn't there.
Vibe coders are measurably more satisfied.
No vibe coding: 5.93/10. Heavy vibe coders: 7.39/10. The gradient is nearly linear. Cross-sectional data, so I can't prove causation. But the correlation is strong enough that dismissing it requires its own explanation.
Quick Hits
State of Prototyping: Spring 2026 — The full interactive report with all 12 sections, methodology, and editor's notes.
Download the raw data — Full de-identified dataset under CC BY 4.0. Run your own analysis.
Open API — Query the data programmatically.
MCP Server — Connect the dataset to your AI tools directly.
That's it for this one. This isn't a weekly issue, it's a snapshot of a moment in time. If something in the data surprises you, I want to hear about it.
See you soon,
Tommy
