Industry
Fintech
Employer
Conta Simples
Time Invested
2 months
How a fixed survey became the engine behind a simplified signup flow
AI Coding, Figma, Client Interviews
How I fixed a broken CES survey and turned it into a recruiting engine for client interviews, batch-inviting respondents over WhatsApp to multiply how many conversations we ran, then used those insights to specify a simplified signup flow.
Outcome
15%
increase in completed signup conversion, in month 1 with the solution live
70%
increase in CES survey responses, versus the period when the survey was broken
~15x
more client interviews per week (from ~1/month to up to 5/week)
Context / Problem
As Product Designer on the Growth / Client Life Cycle squad, while mapping the signup flow I found the CES survey at its end wasn't rendering: we were losing responses. It also measured satisfaction instead of effort, so it wasn't a real CES. I rewrote the question and repurposed the survey as a recruiting tool, asking clients if they'd talk more about their account-opening experience. I automated recruiting by batch-inviting respondents over WhatsApp via Blip: interview volume jumped from 1-2 a month to 3-5 a week, with a waiting list. It became the first case leadership praised as a reference for how designers should work. Those interviews became the basis for simplifying signup.
My Role
I fixed the survey myself, end to end: rewrote the question in Typeform and shipped the implementation, with my own PR tested before production. From the interviews it recruited, I connected the inputs to the simplification thesis the PM was already shaping: which screens were dispensable, what could be pre-filled from bureau data, and where to trim the shareholder registration. I specified the solution from there.
Discovery
The pattern that came out of the interviews: signup asked for too much information, across too many screens, for data the company could already provide or that we could pull from bureaus, open finance, and open banking. From there the thesis became concrete: cut what people need to type, pre-fill what we can already know, and close the flow with a single review screen instead of multiplying steps, applying progressive disclosure.
Solution
The solution followed one principle: fewer fields to fill, as much pre-filled as possible, and a final review screen instead of several data-entry screens (progressive disclosure). From there I designed two flows: an ultra-simplified one for MEI, and another for LTDA and SA with a simplified shareholder registration too.
