UX decisions
The original design starts from the merchant's existing platforms, relationships and language. Its source specifications explain the individual screens; these are the decisions that connect the system.
| Decision | Why it exists | Source |
|---|---|---|
| WhatsApp coexistence before an empty ERP | The shop already has identity, contact names and trading history. Provider-authorized context gives the workspace a meaningful starting point. | History contract |
| Progressive commitments | Joining a list, verifying a phone, completing a profile and authorizing business history each ask for different trust. The later native flow acquires them in that order. | Mobile contract |
| Prep as home | One shop acts on today's work. Prep groups sales by the block where they go, ahead of a general sales dashboard. | Design specification |
| Source conversations beside records | A role or fiado is an interpretation of someone's words. Showing its provenance and reasons makes merchant confirmation meaningful. | Reader |
| Familiar money rails with a durable page | The seller keeps their own receiving destination. The page supplies description, exact amount and persistent settlement identity. | Collection contract |
| Transparent trade indicators | Starting / Steady / Punctual expose counted windows, denominators and evidence floors. They make behavior understandable before any financing claim. | Record arithmetic |
| Native questions sized to the keyboard | One question, a direct input, progress and one forward action stay inside the visual viewport. Auth Tab reuses the browser's actual login session for Meta. | Native UI |
| Explicit action states | A share sheet, accepted Graph message, eligible tester, Play install and bank receipt mean different things. Their labels correspond to their actual source. | Workflow |
The browser retains the original Inter/Fraunces typography, shared theme, artwork and responsive shell. The native bundle finishes at verified onboarding and the waitlist/business pass; its scope differs from the full browser shop.
The historical personal-account experiment ranked referrals by laughs so family chats could exercise the journey. This release selects the already implemented economic evidence rule. The test-only alternate remains documented in history/rank.ts and its explicit replay tests; it supplies useful provenance for why the two rules exist.