Language problems usually start upstream
When bilingual products feel awkward, the issue is often not the translation itself. The issue is that product decisions assumed one language all along: narrow buttons, fixed hierarchy, rushed error states, and labels written before the action was properly understood.
What changes when language enters early
If translation is treated as product work, the team designs differently:
- labels are shorter because intent is clearer
- components leave room for natural phrasing
- edge cases are discovered before they ship
The outcome is not only better Arabic. It is better product thinking overall, because fuzzy decisions are harder to hide.
The bar we want
Users should not feel that one language got the fully considered version and the other received the adapted one. Both should feel authored. That standard requires design, engineering, and language work to stay close together from the start.
