Global Business Account & Spend Management
A product exploration designing a unified financial hub for global startups, closing the gap between multi-currency accounts, international payroll, and employee corporate cards.
↑ Concept: one balance sheet for a startup operating in five currencies at once.
The fragmented finance stack of a global-from-day-one startup
A typical modern startup operates globally almost immediately, incorporated in the US, hiring developers in Poland, selling software to customers in the UK. To manage this, most duct-tape together a traditional bank account, high-fee international wire services, and a separate corporate card provider.
That fragmentation creates constant friction for the CFO: hidden FX markup, slow vendor payments, and days lost every month exporting CSVs to reconcile expenses by hand.
Traditional Bank
Holds USD. Charges 3% FX markup and $45 per international wire. Takes 3 days to settle.
Standalone Corporate Card
Not connected to the main ledger. Requires manual pre-funding. Receipt-chasing happens over Slack.
P2P FX Service
Used strictly for paying international contractors, just to avoid the bank's wire fees.
What this exploration covered
As an independent concept exploration, I owned discovery, information architecture, and interaction design end to end, grounding every decision in practitioner interviews rather than internal assumptions, since there was no live product team to validate against.
CFO & Finance-Lead Research
12 remote interviews with startup CFOs and Heads of Finance on how they actually manage multi-currency operations today.
Information Architecture
Designed the "Base Currency Anchoring" model for the Treasury dashboard, resolving the tension between aggregate and per-currency views.
Spend Control Design
Designed the Self-Serve Spend Flow , merchant-locked, budget-capped virtual cards issued without CFO bottleneck.
Impact Modeling
Modeled time and margin recaptured from automating reconciliation and eliminating bank FX markup.
Three compounding failures in the "Frankenstein" setup
Money is tracked nowhere centrally
Finance leads were manually downloading CSVs from a US bank, a European bank, and three card platforms, then using VLOOKUPs in Excel to reconstruct one picture of company spend.
Corporate cards bottleneck on the CFO
Cards carried enough risk that CFOs hoarded them, so marketing couldn't buy ad spend without physically asking for the plastic card, a workflow that scaled terribly past a handful of employees.
FX costs were invisible until the invoice
A 3% FX markup buried inside a wire fee is easy to miss transaction by transaction, and brutal in aggregate across a full year of international payments.
How might we give a lean finance team one place to receive, hold, convert, and delegate money globally, so month-end close stops being a multi-day forensic exercise?
If we unify receiving, converting, and spending money behind one ledger, with a self-serve, guardrailed spend flow and automatic accounting-software sync, then finance teams will recapture the hours currently lost to manual reconciliation, and stop discovering FX costs after the fact.
The real problem wasn't moving money: it was tracking it
To move past assumptions, I conducted 12 remote interviews with startup CFOs and Heads of Finance. A recurring pattern emerged immediately: "month-end close" was a source of genuine dread. I documented the fragmentation with a detailed journey map, which made the actual UX problem visible, it was never about moving money faster, it was about seeing it in one place.
CFO & Finance Lead Interviews · n=12
Remote sessions surfacing how finance leads actually reconcile spend today, and what "done" looks like at month-end.
Journey Mapping
Mapped the full month-end close journey across three disconnected platforms, exposing exactly where the VLOOKUP workaround entered the process.
What 12 CFO conversations converged on
"Base currency anchoring" is the core IA problem
A CFO needs their total liquidity in one home currency at a glance, while still trusting each individual currency ledger is one click away, not buried, not merged away.
Card-hoarding is a trust problem, not a policy problem
CFOs weren't being controlling for its own sake, they simply had no mechanism to delegate spend safely. Give them that mechanism, and the bottleneck disappears on its own.
Reconciliation pain was a sync problem, not a reporting problem
Finance leads didn't need better reports out of their existing tools, they needed the transaction data to already be correct and matched by the time it reached Xero or QuickBooks.
The rules we designed by, and why
1 · Anchor to one currency, never hide the rest
Derived from interviews showing CFOs need a single liquidity number without losing trust in the underlying ledgers.
→ USD-equivalent total balance surfaced first; each currency's real ledger one click away.
2 · Self-serve spend, with programmatic guardrails
Derived from the card-hoarding pattern. The fix wasn't more permission. It was safer permission.
→ Merchant-locked, budget-capped virtual cards routed through manager approval, not the CFO personally.
3 · Design for the exit, not just the request
Derived from the realization that a delegated card is a liability the moment its owner leaves the company.
→ Instant card pause tied directly to employee offboarding, not a manual finance task.
4 · Automate reconciliation, don't just speed it up
Derived from realizing the VLOOKUP workaround was a symptom. The real fix was correct data at the source, not a faster export.
→ Transactions and matched receipts sync automatically to Xero/QuickBooks, no export step required.
From interview transcripts to a treasury dashboard
01 Iteration 01 · Synthesis Journey Mapping the Month-End Close
I synthesized the 12 CFO interviews into a single journey map of the month-end close, plotting exactly where finance leads dropped into spreadsheets and manual cross-referencing. That map became the brief: solve for the handoffs between systems, not any one system in isolation.
02 Iteration 02 · Information Architecture Resolving Base Currency Anchoring
I explored several dashboard hierarchies before settling on a USD-anchored total balance with individual currency ledgers surfaced as immediate, one-click detail, rejecting both a flat multi-currency list (no sense of aggregate liquidity) and a fully merged single number (hid currency-specific risk).
03 Iteration 03 · Spend Delegation Prototyping the Self-Serve Card Request Flow
I prototyped a mobile card-request flow where an employee names a purpose, sets a budget and frequency, and routes the request to their manager. The approval loop needed to be fast enough that employees would actually use it instead of reverting to expense-and-reimburse , which is why Slack-based approval, not an in-app-only flow, became a deliberate design choice.
The choices that shaped the concept
Decision 01 · USD-anchored dashboard vs. currency-parity view
Tension: Treating every currency equally feels fairer, but buries the one number CFOs actually scan for first.
Choice & trade-off: I anchored to the home currency by default, with per-currency ledgers one click away, optimizing for the CFO's most frequent question ("what do we have?") over theoretical fairness between currencies.
Decision 02 · Self-serve card issuance vs. centralized CFO approval
Tension: Self-serve issuance introduces spend risk that centralized approval avoids entirely.
Choice & trade-off: I chose self-serve, gated by merchant-lock, budget cap, and manager approval, accepting a small increase in process complexity in exchange for removing the CFO bottleneck entirely.
Decision 03 · Real-time ledger sync vs. nightly batch sync
Tension: Real-time sync is meaningfully harder to build reliably than a nightly batch job.
Choice & trade-off: I designed for real-time, batch sync would have just moved the "anxious wait" from month-end to next-morning, not removed it.
One ledger, every currency, no spreadsheets
↑ The Treasury dashboard, USD-anchored total balance up top, individual currency ledgers one click below it.
Request New Card
↑ The Self-Serve Spend Flow, merchant-locked, budget-capped, routed to a manager instead of the CFO personally.
Practitioner input in place of a product team
As an independent exploration, there was no engineering or compliance team to negotiate with, so the 12 CFO interviews had to do double duty as both research and reality check.
Testing feasibility assumptions against real workflows
Every design decision: from Slack-based approvals to real-time ledger sync, was checked against how these finance leads actually worked today, not how a clean-room process might look on a whiteboard.
Being explicit about what's modeled vs. validated
Without a production team, I kept a clear line in my own notes between what interview data actually supported and what remained an informed bet, a discipline that shows up in how the impact metrics are labeled below.
How we arrived at the solution
Card Sorting
Ran unmoderated open card-sorting sessions with SME business owners to see how they categorized multi-currency accounts and payroll features.
Value Proposition Canvas
Mapped the core pain points of SME founders dealing with international suppliers against what the concept could realistically promise.
Illustration Placeholder
Prompt: A stylized illustration of a card-sorting research session for a global business account, international banking concepts on sticky notes, in emerald and silver accent tones. Rendered as a clean illustration on a fully transparent background (no backdrop, scene, or color fill), so it displays cleanly on both light and dark page themes.
Projected outcomes for a lean finance team
This is a concept exploration, not a shipped product. Figures below are modeled from CFO interview data on time spent and fees paid today, not measured production results.
Honest reflections from the process
What Worked
Interviewing practitioners, not personas
Talking to 12 real CFOs surfaced "base currency anchoring" and the VLOOKUP workaround, specifics no amount of internal brainstorming would have produced.
Naming the real problem before designing the dashboard
Recognizing that "tracking money" was the actual problem, not "moving money," reframed every subsequent design decision toward sync and visibility rather than transaction speed.
What I'd Do Differently
Test the self-serve card flow with actual employees, not just CFOs
All 12 interviews were with finance leadership. The requesting employee's side of that approval flow, patience, clarity, friction tolerance, is a different perspective I'd want validated directly.
Pressure-test the real-time sync claim with engineering earlier
Real-time ledger sync across multiple banking rails is a genuinely hard distributed-systems problem. On a real project I'd want an engineering feasibility check in week one, not as an assumption baked into the concept.