Research team

Prototype evolution

A2 lowered word count but increased task time and confusion signals. A3 keeps the helpful facts, moves short cues closer to decisions, and removes result-screen dwell.

Version lanes

vA1

Archived

Verbose clear language

Explanatory panels on most steps, longer result copy, and an animated transfer timeline. Strong cognitive-distance separation, but slower task completion.

Frozen copy — June 2026 friends-and-family pilot

vA2

Archived

Transparent minimal

Same facts and scoring visibility, placed at confirm/result only. Shorter copy, no timeline dwell — targets near-zero distance with faster completion.

June 2026 restart sample; slower than vB in early data

vA3

Archived

Guided minimal

Short just-in-time cues before decisions, compact result confirmation, and no study-task timeline dwell. Targets low cognitive distance with faster completion.

July 2026 cohort before medical referral clarity pass

vA4

Archived

Guided minimal + referral clarity

Same guided-minimal structure as A3, with clearer medical copy that the first visit is with a general doctor — not a dermatologist — and referral comes at that visit.

July 2026 cohort; matched vB on task time

vA5

Live

Answer-shaped guided minimal

Each embedded question's canonical fact appears verbatim as its own scannable line on the screen where it becomes decision-relevant, results lead with the verdict, and filler that only restates visible UI is cut.

New participants are assigned this version

Unchanged

Institutional language

Formal labels, dense terminology, no plain-language explanations. The institutional baseline across all three modules.

Same design since study launch

All prototypes

Three modules × five versions. Flow = interactive preview · Walkthrough = guided replay with prediction questions

vA1

Verbose clear language

vA2

Transparent minimal

vA3

Guided minimal

vA4

Guided minimal + referral clarity

vA5

Answer-shaped guided minimal

Institutional language

Banking

Send money

vA1