When customer behavior starts driving the team’s priorities.
I reoriented a mature but operational design team to turn customer behavior into business decisions — within a regulated, multi-segment FX ecosystem.
Results since 2024, ongoing, as Senior Design Manager.
Many customer profiles, a single regulated FX domain.
Remessa Online is an FX ecosystem for individuals and businesses, with products tailored to each profile. On the B2C side it runs on web and app — and the breadth is deliberate: under the same platform you find the traveler paying for a course abroad, the importer-exporter, the creator cashing out their Google Ads, and the professional receiving a salary in dollars from a company overseas. Customers who look nothing alike, with needs that never repeat.
On the B2B side there is a layer with more weight: FX operations, an exchange API for those whose business is FX itself, and a B2B/B2B2C platform where partners manage their own base. And operating FX in Brazil means operating under the Central Bank’s rulebook — sharper on B2B, where a company may move millions across countries, demanding rigorous analysis on every operation.
The real complexity is not in having many products — it distributes well once you look journey by journey. It is in trust: moving someone’s reais and dollars cannot be an isolated event, at the risk of the platform becoming an FX calculator visited once a month.
Design’s job is the opposite — turning safety and convenience into a relationship asset. The importer-exporter who gets the right template at the right moment, or the B2B customer who generates an invoice already filled with Remessa’s details — one of the site’s most-used features — show how the journey shortens the path between intent and operation, and sustains recurrence.
Insight wasn’t missing. Turning it into decisions was.
When I took over, the team was shipping a lot: dense Figma files, a visible volume of screens. But the work was mostly operational — poor handoffs, and a development team that sometimes didn’t even open Figma. Design rushed to the next task, and at that pace its own voice had weakened, even among peers. Screens were made; decisions were not contested.
The strongest signal came from an almost ironic counterpoint. The research front was mature — a reference researcher ran strategic studies, cross-referencing quant and qual with access to hard-to-reach profiles. The insights existed, and they were good. But they stopped at the reading: they became knowledge and presentation, not movement. They didn’t trigger a workgroup on churn, an investigation, a feature.
The central challenge wasn’t to make the team ship more. It was to raise the bar of a team that delivers craft and reads the business — that turns customer behavior into decisions, without losing storytelling or interface along the way.
And there was a structural backdrop: leading the turn with a lean structure, from a leadership seat under the Product pillar, without the formal authority of a Head. The change would have to come through influence and reference, not mandate.
Filter what comes in, find the signal at the source, communicate to influence.
The first move was diagnostic. In the first 90 days, I mapped the challenges — people, business and platforms — alongside 1:1s, a kickoff and a workshop retrospective, to capture pain points, skill gaps and feed the development plans.
On that base, what the team looked at — and how — changed. Designers began running strategic research and looking at the customer’s whole journey, not just the screen flow. And they started seeking the signal at the source, with autonomy: data straight from CX, sales and marketing, cross-referenced with the business indicators — operations, spread, average ticket. Design moved out of dependence on the PM and started reaching other areas on its own.
Craft and communication aren’t finishing touches — they’re what gives design a voice before the business.
Autonomy came through capability transfer: research standards, templates and teaching material, sustained by 1:1s and development plans on a cadence. That is how running discovery stopped depending on me and became the team’s routine.
When behavior becomes decision, and decision becomes result.
The first result was operational. The team went from 6–10 operational deliveries per sprint to 3–4 with discovery built in, and began running 2–3 discoveries per month — from research to handoff in about a month. Recurring research became an engine: 5–6 studies per month (CSAT, CES, churn) plus a quarterly NPS, feeding decisions continuously.
That engine started moving specific profiles. For the investor, screen recordings and funnel analysis exposed a mismatch between the platform and the brokerage in the send flow. Once nomenclature, data and triggers were aligned, the profile’s web conversion rose from 20–30% to 47%. In COMEX, reading behavior revealed churn tied to logistics and documentation; we refined the journey to reduce that friction — a front still evolving.
On the business side, the work connected to revenue and efficiency across the fiscal cycle: triggers like Pix settlement deadlines, migration from TED to Pix, faster settlement. The average ticket for the quarter nearly doubled — via a cross-functional workgroup with product, sales and marketing, but started from design’s research. And trust became part of the interface: clear operation status, security and data under Brazil’s LGPD.
In FX, convenience opens the door — but it’s trust that sustains recurrence.