I started by mapping the ex-Ituber's full journey with the team, looking for where the pain actually lived, not where we assumed it did. Support-ticket data pointed straight to the real bottlenecks, and from there we prioritized with an effort x impact matrix, choosing the simple fixes that solved the most. I loved this part especially, it's rare to have data this clear steering every design call. I designed the target flow and built the prototypes inside the iDS, Itaú's design system, which kept design and development speaking the same language from day one. I tested it with 8 real users using SUM, and the results gave me the confidence to move forward on evidence instead of opinion. Handoff happened in a straightforward conversation with the dev team, walking through the flow in Figma and explaining how the iDS pieces fit together. Having everything pre-coded sped up implementation and kept the experience consistent end to end.