Each AI setup was given the same brief, word for word. It wrote a brand plan, then built a style guide, a landing page and a dashboard, then reviewed and revised its own work: take 1 is the first build and take 2 the revised one. The letters are random and unrelated to the letters in brief 1. Tap a screenshot for the full page, or open the live pages.
Which one would you pick, and why? Any thoughts on what works or doesn't are welcome. Your pick is saved in this browser only.
We're launching a reservations and appointments app for barista-catered events and coffee experiences: customers book a barista (and a cart or bar setup) for events, offices and private tastings, and manage their appointments. Plan first, then build. Nobody is available to answer questions, so make sensible assumptions and write them down.
Planning: write PLAN.md covering the brand (name if needed, positioning, audience), brand elements (logo or wordmark direction, colour, type, imagery, iconography, motion principles), hierarchy and information structure for the landing page and the app, and brand personality with sample copy (voice, tone, headlines, microcopy).
Then build three deliverables that share one design system:
1. ./style-guide: a design system and style guide page (tokens, colour, type scale, spacing, components, motion, voice and copy examples).
2. ./landing: a landing page that showcases and sells the app, with motion graphics and interactive elements (for example an animated product demo or an interactive booking preview), responsive, with reduced-motion support.
3. ./dashboard: a shadcn dashboard for the business side (upcoming bookings, a calendar or schedule view, revenue and utilisation stats, barista roster), in the existing shadcn project, themed to the brand.
Each setup ran this brief once, then was asked in the same session to review its own work as a senior brand and product designer, write REVIEW.md, fix the problems and write CHANGES.md. Take 1 is the first build; take 2 is the same work after that review and revision, so compare them to see how much the revision helped. Automated checks: builds; axe-core 4.14 (WCAG 2.2 AA rules) on each page at 1440 after scrolling through; sideways overflow at 390px; the landing page scrolled with reduced motion off and on (how many animations ran, endless ones still running, text still invisible after scrolling); and a shadcn style scan of the dashboard source. Numbers in axe cells are affected elements (rules); hover for rule names.
| Take | Build: style guide | Build: landing | Build: dashboard | axe: landing | axe: dashboard | axe: style guide | Overflow 390: style guide | Overflow 390: landing | Overflow 390: dashboard | Landing animations while scrolling (normal / reduced motion) | Endless animations left with reduced motion | Text left hidden with reduced motion | Console errors: landing | Console errors: dashboard | Dashboard: raw palette colours | Dashboard: hard-coded colours | Dashboard: space-x/y | Dashboard: dark: overrides | Dashboard ui components: CLI / hand-written |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| G1 first build | static | static | pass | 1 (1) | 7 (1) | 7 (1) | 0 | 0 | 298px | 16 / 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 7 / 0 |
| G2 after revision | static | static | pass | 1 (1) | 0 | 7 (1) | 0 | 0 | 0 | 15 / 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 1 | 7 / 0 |
| K1 first build | static | static | pass | 0 | 3 (1) | 0 | 0 | 0 | 0 | 4 / 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 13 / 0 |
| K2 after revision | static | static | pass | 0 | 3 (1) | 0 | 0 | 0 | 0 | 4 / 2 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 13 / 0 |
| Q1 first build | static | static | pass | 3 (1) | 5 (1) | 3 (1) | 0 | 0 | 85px | 22 / 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 7 / 0 |
| Q2 after revision | static | static | pass | 0 | 7 (1) | 3 (1) | 77px | 0 | 85px | 22 / 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 7 / 0 |
| R1 first build | static | static | pass | 0 | 4 (1) | 1 (1) | 0 | 0 | 0 | 33 / 0 | 0 | 0 | 0 | 0 | 0 | 4 | 0 | 2 | 12 / 0 |
| R2 after revision | static | static | pass | 0 | 4 (1) | 1 (1) | 0 | 0 | 0 | 33 / 0 | 0 | 0 | 0 | 0 | 0 | 4 | 0 | 6 | 12 / 0 |
| V1 first build | static | static | pass | 0 | 5 (2) | 1 (1) | 0 | 0 | 0 | 18 / 6 | 0 | 12 | 0 | 0 | 0 | 0 | 0 | 0 | 12 / 0 |
| V2 after revision | static | static | pass | 0 | 5 (2) | 1 (1) | 0 | 0 | 0 | 18 / 7 | 0 | 12 | 0 | 0 | 0 | 0 | 0 | 1 | 12 / 0 |
| W1 first build | static | static | pass | 0 | 8 (1) | 0 | 0 | 0 | 0 | 7 / 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 7 / 0 |
| W2 after revision | static | static | pass | 0 | 0 | 0 | 0 | 0 | 0 | 7 / 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 7 / 0 |