A product designer with 8 years in Australian tech watched PMs ship AI-generated prototypes straight to production — no design review, no interaction states, no structural decisions.
The workflow: PMs use AI to prototype features, run weekly user tests, then hand prototypes to developers with instructions to “slap on an open-source design system.” The designer, who also uses Claude and Cursor, does the same — but catches what PMs miss.
The gaps the poster cataloged:
- Interaction states — hover, focus, active, disabled: missing or inconsistent across shipped features
- New-user usability — onboarding and first-run experience unconsidered in AI-generated flows
- Design structure — no component architecture, no token system, no hierarchy beyond what the AI guessed
- Information density — progressive disclosure ignored; screens dump everything at once
- “Research” docs — hundreds of AI-generated documents, dense text, zero substance, and employees are expected to reference them
The design debt is already cascading. What ships as a “working prototype” lacks the interaction states, progressive disclosure, and structural decisions that separate a demo from a product. The poster notes these gaps seem irrelevant at the PM level — but they surface in production as usability failures and rework.
NOTE
The Reddit post cuts off mid-sentence: “Rather optimistic about the fact that these AI artef…” — the full reasoning is unknown. The designer sees the debt accumulating but isn’t sounding a full alarm.
The pattern: non-designers using AI to skip the design process, then patching aesthetics with a component library. The bill comes later — in onboarding drop-off, support tickets, and structural rework that costs more than doing it right the first time.