mh studio Book a free call
← All work
Finance · Cross-border paymentsFictional client

Corridor: know what lands before you send.

A mobile and web app for remote freelancers who get paid in dollars by clients abroad and send part of it home to family.

Try the live prototypes ↓
Role
Lead product designer
Deliverables
Website, iOS/Android app, web app, design system
Benchmarks
Wise, Remitly
Stage
Seed, pre-launch MVP
Team
Solo concept (assumed: PM, 2 engineers)
Try the product

Don't just read it. Use it.

Every screen below is a working prototype. Send money, track it live, flip the card, explore the dashboard.

corridor.

Welcome back, Emily

Locked for your safety

Mobile app

14 working screens. Everything is connected: send, convert or add a recipient and the balances, history and insights update.

  1. 1Tap Unlock with Face ID, then hide balances with the eye icon.
  2. 2Tap the center Send button, type on the keypad, pick Sarah and hold to send.
  3. 3Open Convert, swap currencies and watch both balances update.
  4. 4In Activity, switch to Insights and tap the months. Tap any payment for its receipt.
  5. 5On Rates, drag the alert and switch 1W, 1M and 3M.
  6. 6In Profile, add a new recipient, then send them money.
Open full screen ↗

The brief

Corridor's founders came with a clear goal: build the first version of a money app their target users would trust with their income on day one.

Their users juggle three jobs: getting paid by foreign clients, converting at a fair rate, and sending money to family. Today that means a mix of platforms, bank transfers and guesswork about what will actually arrive.

Payments is the biggest fintech category, about 44% of fintech revenue (BCG Global Fintech Report 2026), so the bar set by the leaders is high.

Concept project: Corridor and its founders are fictional. The brief, team and constraints below are assumptions I set to make the work realistic, and they are labelled that way.

00 · My role and the setup

What I owned, and the box I worked in.

What I owned

  • Product framing: personas, jobs to be done and the v1 scope cut
  • Benchmark teardown of Wise and Remitly
  • Information architecture for mobile, web app and website
  • Every screen, including the failure and edge states below
  • The design system: tokens, type, components and accessibility rules
  • The three working prototypes on this page, built in code

Assumed team

Concept framing, not a real engagement

  • A founder acting as PM, who owns pricing and partners
  • Two engineers, mobile and web, reviewing feasibility weekly
  • A part-time compliance advisor who signs off verification and hold copy
  • Legal copy on the website is a placeholder for counsel to review

Constraints I designed to

Assumed, the way a seed-stage payments team would face them

  • Regulated money flows: identity checks before the first send, and some transfers get held for review
  • One corridor at launch: USD in, EUR out
  • A rate lock only as long as the FX partner will honour it
  • Two engineers, so one component library across mobile and web

Planned timeline for this concept

About 7 weeks, split into five phases. This is how I would plan the work as a real engagement, not a record of time spent.

  1. Week 1
    Discover
    Wise and Remitly teardown, personas and the v1 scope cut
  2. Weeks 2 to 3
    Design
    Information architecture and the send, receive and card flows
  3. Weeks 4 to 5
    Prototype
    Working prototypes of the mobile app, web app and website
  4. Week 6
    Build
    Design system tokens, components and developer handoff
  5. Week 7
    Test and launch
    Usability tests on the first transfer, fixes and launch checklist

01 · Who we designed for

Primary persona (assumed, to validate)

Emily, 27

Freelance illustrator, paid in USD by 3 to 5 clients

  • Wants to know the exact EUR amount before converting
  • Sends money to her mother every month
  • Needs proof of income for rent and visas
Jobs to be done
  1. Get paid: share USD account details with a client
  2. Decide: see the amount received, fee and timing
  3. Send: pay a saved recipient in under a minute
  4. Relax: track the money until it arrives

02 · Benchmarking the leaders

What we kept, and what we refused to copy.

Kept

  • A live quote calculator as the core interaction
  • Rate lock with a clear time window
  • Fee shown as one line, before payment
  • Step-by-step transfer tracking

Refused

  • Eight competing calls to action on one screen
  • Key terms hidden in footnotes
  • Personal and business products mixed together
  • Long, unsearchable lists of currencies

From a section-by-section audit of the Wise US homepage, November 2026.

03 · The core flow

Five screens from home to "it arrived".

04 · Key decisions

Six calls, and what I said no to.

Each decision shows the option I rejected and why. The trade-offs are reasoned from the benchmark and the constraints above, not from user testing.

Lead with what lands

The EUR amount the recipient gets is the largest number on the amount screen. The USD you send is the input, not the headline.

RejectedSend amount as the headline, the way most banking apps do it. People anchor on the big number, and the question they actually have is "what will Mum get?"

One primary action per screen

Continue, then Confirm and send. Secondary options live in clearly labelled text buttons, never as a second filled button.

RejectedQuick actions (convert, schedule, split) on the review card. They would save a tap for power users but make the one step that moves money harder to scan.

Terms next to claims

Rate lock window, fee and delivery time sit on the same card as the amount. No asterisks.

RejectedA "Fees and terms" link under the button. Cleaner, but it is exactly the footnote pattern the benchmark audit flagged as costing trust.

Hold to send

The final step is a press-and-hold with a progress fill, so nothing is sent by a stray tap in a pocket or on a bus.

RejectedA separate confirmation screen. It is safer than one tap but adds a full screen to every repeat send. Cost I accepted: hold is slower for power users, so I would revisit it if repeat senders drop off at this step.

Verify only when it matters

Anyone can browse rates and get a quote without ID. Identity is checked once, right before the first send, with a 3-step progress bar and a plain "why we ask" line.

RejectedID upload at sign-up. It front-loads the hardest step before people have seen any value, which is where fintech onboarding usually leaks.

Two lock windows, on purpose

Conversions between your own balances lock for 30 seconds because they execute instantly. Transfers show a lock countdown, because they settle later. The prototype shows 24 hours; I assumed an FX partner absorbs that risk.

RejectedOne long lock for everything. If treasury can only fund a 2-hour window, the same countdown handles it, and the expiry state (below) requotes in plain words instead of failing at the last tap.

Open the web app → Open the website → Open the design system →

Mobile is for sending. The web app is for the monthly money admin: account details for clients, history, statements and rate alerts.

05 · Edge cases and failure states

Money apps earn trust in the bad moments.

The happy path is five screens. These are the states I expect people to hit most, designed to one rule: say what happened, what we are doing about it, and what you can do, with a time estimate where there is one. Copy for holds and verification is written for compliance sign-off.

ID verification

  • What happened: You are about to send for the first time, so we need to know who you are.
  • What we're doing: We check your ID once. Most checks finish in a few minutes.
  • What you can do: Take a photo now, or keep browsing and come back. Your quote stays saved.

Recipient name mismatch

  • What happened: The name you typed is close to, but not the same as, the name at Sarah's bank.
  • What we're doing: We show both names side by side before you send. "Matches" and "Doesn't match" are the other two states.
  • What you can do: Use the bank's name in one tap, or edit the details.

Compliance hold

  • What happened: Your transfer was picked for a routine check. This is normal for some first or larger payments.
  • What we're doing: We tell you how long it usually takes and keep your rate locked while we check.
  • What you can do: Upload the invoice to speed it up. We notify you the moment it moves.

Returned transfer

  • What happened: Sarah's bank sent the money back because the account details did not match.
  • What we're doing: The full amount, fee included, is back in your USD balance. No one is left guessing.
  • What you can do: Fix the details and resend with one tap, using the same amount.

Fee transparency

  • What happened: Nothing went wrong. This is the card that prevents most "where did my money go?" questions.
  • What we're doing: Fee, rate, lock window and amount received sit on one card, worked out in front of you.
  • What you can do: See the full sum before you hold to send. No asterisks.

Rate lock expired

  • What happened: You took longer than the lock window, and the market moved.
  • What we're doing: We requote straight away and show the difference in euros, not just a new rate.
  • What you can do: Send at the new rate, or go back and change the amount.

Edge-case inventory

Already working in the prototype above: live IBAN check when adding a recipient ("Looks like a valid IBAN"), a disabled "Not enough balance" state on Continue, name check on the review card, card freeze, and a hold-to-send that cannot fire on a tap. Next on the list, not designed yet: delayed transfer, cancel before conversion, recall after conversion, and daily limit reached.

06 · Accessibility decisions

Built for people who send money tired, in a hurry, or by ear.

Hold to send has a non-hold path

The hold works with Space or Enter held down, not only a finger. For screen reader and switch users, the spec swaps it for a standard "Confirm and send" action, because a timed hold is hard to do with assistive tech.

Status is read out, not just shown

Tracking headlines, toasts and the lock screen message are live regions, so "Arrived" or "Held for a check" is announced without hunting for it.

Never colour alone

Every state pairs colour with a symbol and a word: ✓ Matches, ! Close match, ↩ Returned. Saffron is never used for text on light backgrounds, because it fails contrast there.

Big enough to tap, clear enough to read

Touch targets are at least 44px. Text passes WCAG AA contrast (4.5:1). Amounts use figures of equal width so columns of money line up.

Motion is optional

Count-ups, the Face ID scan and screen transitions switch off when the system asks for reduced motion. Nothing important is only communicated by animation.

Plain words for money

"Sarah gets", not "Beneficiary receives". Hold and verification copy avoids legal terms, because the people most likely to hit them are first-time senders.

07 · Design system

Calm, warm and unmistakably about money.

Deep Teal #0F3D3E

Saffron #FFB547

Paper #F6F4EE

Ink #0B1414

Success #1F8A5B

Aa

Manrope across the product. Numbers use equal-width figures so amounts line up in columns. Amounts use the heaviest weight (800), body text a regular one (500).

08 · Success metrics (targets)

How I'd measure success at launch.

Targets, not results. This is a concept, so there is no live data. These are the 90-day targets I would propose to the team, with the event that measures each one.

Quote to send

35%+

Share of quotes that become a confirmed transfer.

quote_viewed → transfer_confirmed

Time to send

Under 30 s

From tapping Send to confirming, for a saved recipient.

send_started → transfer_confirmed

"Where's my money?" tickets

Under 5

Support contacts per 1,000 transfers. Tracking and the failure states should push this down.

support_ticket{topic: transfer_status}

Repeat sends

40%+

Users who send again within 30 days.

transfer_confirmed (2nd within 30d)

ID check completion

80%+

Started verifications finished in the same session.

kyc_started → kyc_submitted

09 · What I left out of v1

  • More currency corridors. One pair (USD to EUR) keeps rates, compliance and support simple while the core flow proves itself.
  • Business accounts. The benchmark showed what mixing personal and business does to clarity. Freelancers first.
  • Scheduled and split payments. Monthly repeat covers the main habit. The rest waits for real usage data.
  • Home screen widgets and live tracking on the lock screen. A natural next step for "track until it arrives", not a launch need.
  • Dark mode. Tokens are named so it can be added later without a redesign.

10 · Reflection

What I would do differently

  • Start with the failure states instead of ending with them. They shape the review card more than the happy path does, so they should lead the work.
  • Pressure-test the 24-hour rate lock with a payments partner before designing around it. It is a promise that costs real money.
  • Prototype the ID check inside the send flow earlier, since it is the step most likely to lose people.

What I'd test first with real users

  • Do people trust the amount received more than the amount sent?
  • Do the hold and delay messages calm people down, or prompt a support ticket?
  • How many people drop off at the ID check, and at which step?

Hiring a Lead designer, or building something like Corridor?

Open to full-time Lead roles and contract work. Remote worldwide.

Email me → See my CV
Next project
Stackwise
View case study →