FXpress · AI Co-Pilot for Cross-Border Remittance · Post-Test Synthesis
🟢 What this debrief covers: Findings from a formative usability study on an interactive prototype of the core interactions — reading the rate, responding to Fexi, setting a trigger, confirming the recipient amount. Every interaction tested is one the deployed widget ships.
| Property | Value |
|---|---|
| 🏷️ Status | Complete |
| 🔬 Source | Formative usability study — 5 participants (Janice, Jerry, Gabriela, Jin, Ollie) |
| 🔖 Format | 7 core tasks · time-boxed · prototype task test |
| ➡️ Feeds Into | Design Iterations · Final Design |
First read of the rate (5/5). All five participants located Fexi's rate status card without instruction or onboarding prompts. Janice: "I actually feel like I understand what this does before I've used it."
Send Now path (5/5). All five completed the primary send flow error-free. The Fexi card lowered the cognitive barrier to initiating a transfer — instead of navigating menus, users responded to a plain-language read. Jin: "This is exactly what Coinbase should have built."
Confirmation resolved uncertainty (5/5). Leading with the recipient amount in local currency resolved pre-transfer anxiety across every corridor tested — ₱11,200 GCash, ₩950,000 KRW, $1,350 CAD, and M-Pesa Instant Payout. No participant asked "what will my recipient get?"
Crypto abstraction (5/5). Zero participants in the default flow asked about, or realized, that stablecoin rails were in use. Jerry — the most crypto-averse participant — asked zero questions about the backend.
Non-predictive copy landed positively. Gabriela responded favorably to the descriptive framing ("Fexi describes…") unprompted, confirming the governance constraint reads as trustworthy rather than evasive.
Trigger vs. Send Now disambiguation (4/5 completed). Equal CTA visual weights caused misclicks. Jerry tapped the primary "Send Now" button while attempting the trigger task, noticed the state change, backtracked, and selected the trigger button. The pattern indicates the two paths need clearer visual differentiation — not a different hierarchy.
Trigger form labels (4/5 completed · 1 abandoned). "Threshold Amount" without a currency suffix caused hesitation and one abandonment. Janice reached the trigger setup screen and abandoned at the unlabeled field. Gabriela hesitated and asked: "I don't know what 1,380 means — is that dollars, won, or shillings?" Jin, a power user, completed it but flagged the "Use This Trigger" CTA label as vague.
Five real participants, each recruited to sit at a distinct point on the Crypto Comfort × Automation Preference spectrum. Task outcomes are reported per participant, per block.