FXpress · AI Co-Pilot for Cross-Border Remittance · Brief
People send money across borders on a schedule — to a parent, a sibling, a child studying abroad. The transfer is routine; the stakes are personal. Existing apps move the money reliably but offer no help with timing, and timing is where senders quietly lose value transfer after transfer.
FXpress is a companion app, not a bank. It sits over the user's existing funding source (their bank, their wallet) and helps them send money across borders at the right time, with confidence in what arrives. Its co-pilot, Fexi, reads the live rate, describes where it sits across the last 30 days of central-bank data, and then waits for the sender to decide. It never forecasts, and it never sends on its own.
Live working prototype: fxpress.netlify.app
This case study is a conceptual product extended into a working artifact. The live widget runs on real FX data via Frankfurter (sourced from 84 central banks), narrates with Claude via a Cloudflare Worker, and computes conversion math deterministically in JavaScript. Production-readiness (server-side rate watching, real notifications, licensed payout integration) is scoped as next-stage work.
Design an AI-assisted cross-border remittance experience that helps users send money internationally at the right moment, with confidence in what their recipient receives. The goal is a trust-driven interface that lets users:
The principle behind the design: code computes; AI narrates; the human authorizes. Every number Fexi is allowed to say comes from code; the language model only phrases the situation warmly, and is barred from claiming anything about what the rate will do next.