Small business owners have been running multimillion-dollar companies on a bank balance and a gut feeling. We found the tools they rely on show them their money — but never what to do about it.

Before touching a screen, we had to earn fluency in a domain most designers never meet: trade credit. The team ran a compressed discovery cycle with the founders to align scope, sequence, and the regulatory guardrails that shape every recommendation surface in a financial product.

1. Domain immersion — trade credit, factoring mechanics, credit insurance, and the C&I (credits & incentives) landscape.
2. Persona and scenario definition, pressure-tested against the four owner archetypes below.
3. Flow architecture and rapid prototyping of the four flagship strategy plays.
4. Trust, disclosure, and compliance review baked into every recommendation — not bolted on before launch.

The “before” state isn’t a legacy portal — it’s a habit. As one engineering-firm CFO put it, the worst pattern in small business is checking the bank balance every day and calling it cash management. The redesign target was the ritual he prescribes instead: update the books, then review weekly — change in cash, A/R balance and aging, A/P balance and upcoming dues.

Old: a bank app, a spreadsheet, a QuickBooks tab, and anxiety.

New: one home screen answering three questions at a glance — what’s my expected net over 90 days, what’s late, and what should I do about it?

We mapped the landscape owners already use: bookkeeping suites that record the past, bank dashboards that display the present, credit bureaus that sell reports, and lender marketplaces that start at “how much do you want to borrow?”

Each covers one link in the chain. None closes the loop from insight to executed strategy — spotting the concentration risk, explaining the fix, and letting the owner act on it in the same session. That loop is Wirth’s lane, and it dictated the product’s center of gravity: the Co-pilot conversation, not the dashboard.

Each persona’s strategy became a decision-support flow with the same skeleton: trigger → explanation → dollar math → action. A concentration alert fires from Noah’s actual receivables; the screen explains the exposure in plain words; the math shows cash-in-hand date and the effective annualized cost of that 3% discount; the action executes without leaving the app. Roger’s insurance flow expresses coverage and the 0.02% monthly premium against his real covered A/R. John’s factoring flow works at invoice level — advance rate, fees, net proceeds. Emily’s monitoring flow ranks customers by risk and pairs every alert with a suggested next move.

Interface principles carried across all four:

• Show the why. Every Co-pilot claim taps through to the underlying invoices and transactions.
• Honest uncertainty. Forecasts render as ranges, never false-precision single numbers.
• Freshness is a feature. Sync stamps (“Updated 12h ago”) are prominent, and degraded-data states are designed, not improvised — a forecast on a broken feed is worse than none.
• Everyday language, on demand depth. Inline glossary for factoring, AVP, letters of credit; severity-tiered alerts so a tight quarter reads as guidance, not 2 a.m. anxiety spam.
• The whole finance triad. Invited, permission-scoped roles for the bookkeeper and accountant — because the person updating the books weekly is often not the owner.