Zum Hauptinhalt springen

Worked Example: Compare Current and Proposed UI Screens

Use a repeatable structure for comparing current and proposed interface screenshots and turning the comparison into one recommendation.

multimodal-ai5 min read

You have current and proposed checkout screens for desktop and mobile and need one recommendation about CTA clarity and trust. Label versions -> Name criteria -> Inspect deltas -> Recommend The usual trap is to ask for “feedback on both screens” and then drown in differences that are visually real but strategically minor. Without named criteria, the model may over-index on aesthetics or novelty instead of the decision at hand. Label Name the four images by role: current desktop, proposed desktop, current mobile, proposed mobile. This prevents the model from losing track of baseline versus variant and device context. Criteria…

Read the full lesson

Sign up free: one personalized lesson every day, matched to your role and goals.

← Back to library

Kontakt aufnehmen