I don't follow a process. I have principles.

I connect design to business outcomes. And I build the team and the systems so it doesn’t depend on me.

Principles

What never changes.

The route changes with every initiative — I start with research when the problem is fuzzy, with a prototype when the solution is the fastest hypothesis. What never changes is the standard I hold the result to.

  • Empathy with data.

    Qualitative + quantitative research

    I talk to users — and read the dashboard, the NPS, the session replay. Intuition isn't guesswork: it's a model built through real observation, not assumption.

    At Remessa Online, reading customer behavior before prioritizing took investor web conversion from 20–30% to 47%.

  • Solution as hypothesis.

    Fast prototyping as research

    Sometimes the fastest way to understand the problem is to build the solution. A working prototype in two days teaches more than two weeks of polished briefs.

    At meutudo., DesignOps plus prototyping shortened time-to-market by 29%.

  • Craft as the edge.

    Quality as intent, not accident

    When AI generates interfaces in seconds, the detail that separates good from great becomes more visible, not less. Timing, hierarchy, micro-interaction — that doesn't come from a prompt.

    At the Tomorrow Lab (Instituto Atlântico), attention to detail sustained ~90% CSAT.

  • Business as the compass.

    Design with an explicit OKR

    Every design decision I defend has a metric attached — conversion, activation, NPS, churn. Connecting experience to outcome is what earns design a seat at the decisions that matter.

    At meutudo., treating the Design System as an internal product drove +36% operational efficiency.

AI + Design

AI compresses time. Not judgment.

With AI, the steps compress — a PM can prototype in code before I've finished the brief. That doesn't destroy the process; it repositions where the designer adds the most value. The right answer isn't to master more tools. It's to have better judgment about which to use, when to stop, and what to discard.

  • Context is the new skill.

    The quality of AI output is a direct function of who can articulate the problem precisely. Prompt engineering is applied design thinking.

  • Prototype first, research faster.

    I use AI to generate prototypes in hours — and put them in front of users before the competitor has finished discovery.

  • The design-to-code gap is closing.

    Connecting design tokens directly to code removes the translation layer where design intent has historically been lost.

  • Craft matters more, not less.

    AI has no taste. It doesn't feel when a transition is wrong by 80ms. It doesn't know that spacing breaks the rhythm. That's still human work — and will be for a long time.

In practice

I adapt the method. Not the standard.

Maturity in design is knowing what to skip — and being confident enough to justify it. Here's what's non-negotiable and what depends on context.

Always
  • Align the success metric before starting Without it, any solution can look right.
  • Validate with real people at some point Could be 3 users. Could be an A/B test. But there is always something.
  • Defend craft in the details to the end I'll concede on deadline. On quality, we negotiate.
  • Revisit metrics after launch Shipping is the start of validation, not the end.
  • Document the decision, not just the result The why protects the team in future sprints.
When it makes sense
  • Full Double Diamond For genuinely fuzzy problems, with broad scope and real time.
  • 5-day Design Sprint I run it in 3 when the team is senior and the problem is well defined.
  • Detailed journey map When the problem spans multiple touchpoints or is new to the team.
  • Formal personas Useful for onboarding stakeholders. Skippable when the team already knows the user.
  • Whole-team HMW sessions Valuable for alignment. Skippable when the direction is clear and the deadline is tight.
The system

A principle that doesn’t become a team isn’t worth much.

The four principles don’t stop at talk — they become the day-to-day architecture.

Not theory — it’s what I structured at meutudo. (org from scratch) and at Remessa.

See it in the cases
  1. Intake Two fronts: a Product + Design backlog and continuous discovery driven by Research.
  2. Backbone Research + DesignOps hold the flow — recurring discovery becomes priority; ops becomes structure.
  3. Rituals · at the right cadence
Next step

Have a product challenge?

If what you just read speaks to your situation — this is the right conversation.

Let’s talk