Making merchant data trustworthy before making it intelligent.

I was brought in to design a few merchant-facing dashboard cards. But every card raised a product question — what does this metric mean, what time range does it cover, can a merchant reconcile it with the money that actually hits their bank? In payments, a dashboard card isn’t a visualization; it’s a trust contract. So I wrote the first PRD myself, and the work grew from a UI task into owning the product.

From November 2025 we put the app in front of partners and tested it. Across several interview rounds with our three core partners, one pattern held: merchants don’t reconcile by “business growth.” They reconcile by sales, tips, tax, fees, and what actually lands in the bank — and their real anxiety is “does what I see match my deposit?” A market signal sealed it: demographic “growth analytics” had under 1-in-10 paid retention in the field. So I repositioned the product from “See your business grow” to “Where does my money come from?”

Before adding features, I defined the data-truth layer. The clearest example: merchants reconcile against bank settlement, which lags across days and months, so Day / Week / Month couldn’t be loose UI labels — they became product definitions, and I separated the selected range from data freshness (“last updated”). I sequenced everything else as a staged roadmap, prioritizing by what has to be true before a feature is safe rather than what’s loudest.

We validated through a POC with our partner P5K, then launched the merchant app across iOS, Android, and tablet, with onboarding and growth ops driving adoption. The result: 3× growth in two months — and noticeably fewer inbound support calls as the data became self-explanatory.

With the data trustworthy, AI became the next layer. I designed two workflows — Discover (surface the anomalies worth attention) and Resolve (intent-driven action, e.g. finding a transaction to refund) — under firm guardrails: no silent execution, status-aware eligibility, and a full audit trail. The human stays the decision-maker on anything that moves money. This is designed and prototyped, not yet shipped.