Beginner · Workflow

Concept visual explore

Mood and brand exploration with a hard decide gate — not an endless Midjourney scroll.

Purpose

Agency and brand cases (e.g. Currys on Firefly) show image gen speeds concepting when teams keep human art direction. Beginners fail by treating the feed as the deliverable. This workflow forces a brief, a small explore set, and an explicit pick before any UI structure work.

Steps

1. **Brief the look** — audience, emotion, three adjectives, one “never” (e.g. no purple glow). No tool yet. 2. **Generate a capped set** — 6–12 stills max across one or two tools (Firefly / Midjourney). Same seed brief. 3. **Cluster** — group by direction; kill clusters that violate brand or accessibility contrast needs. 4. **Decide** — pick one direction; write reject reasons for the others. Export only the winner + one runner-up. 5. **Handoff note** — one paragraph for the UI designer: palette hints, texture, what must not enter the product chrome.

When to use / Skip

Use for early brand/mood before wireframes. Skip when you already have a locked DS visual language — then go to `narrow-canvas-assist` or medium prototype slots. Not for production UI chrome or final marketing assets without legal/IP review (`ethics-as-ops-gate`).