AI-assisted Product Design

I use AI to move from hypothesis to functional product.

Compass UOL · Casas Bahia Pay

A practice for shortening exploration, prototyping and the handoff to engineering without skipping research, judgement or human decisions.

Production case

ajudaQi: a complete redesign for an internal support experience.

Figma, Claude and Figma Make. Validation with support operations and handoff to Engineering through Figma MCP.

The redesign sat inside a backoffice modernisation migrating from Vue to Next.js. The Foundation, Core UMS and User Profile tracks are complete.

Refactoring supported by React and ShadCN. Backend limitations outside scope were documented as future opportunities.

Visual evidence

One operational touchpoint, before and after the redesign.

Before · ajudaQi, previous interface
Previous ajudaQi interface
After · ajudaPay, redesigned interface
Redesigned ajudaPay interface

Project screens used as visual evidence.

The project workflow

Where each tool comes in, and where the decision stays human.

  1. 01Figma

    Exploration and interface

  2. 02Claude · Figma Make

    Alternatives and prototype

  3. 03Support operations

    Validation with the people using it

  4. 04Figma MCP

    Handoff to engineering

  5. 05Engineering

    Refinement and scope

  6. 06React · ShadCN

    Refactoring in production

The process

AI speeds up alternatives. Product judgement remains human.

  1. 01

    Frame

    Problem, context and constraints before interface.

  2. 02

    Explore

    AI accelerates flow alternatives and prototypes.

  3. 03

    Validate

    Operations and technology assess the proposal.

  4. 04

    Deliver

    Design-to-Code and structured handoff.

Technical boundary and human decision

When the backend could not support the desired experience, the limitation was recorded rather than hidden in the prototype.

AI can propose, structure and accelerate. People assess evidence, impact, technical constraints and what should actually reach production.

Where the time changed

Four activities from my own design workflow that now take a fraction of the time.

  • Interview synthesis and interface reading2 to 3 hoursaround 30 minutes
  • Inspecting and reading Figma files3 hoursaround 30 minutes
  • Token mapping and handoff prep4 to 6 hoursaround 1 hour
  • From Figma to a React component structure1 to 2 days2 to 4 hours

Times observed in selected initiatives, not a controlled measurement.

Adoption within Design

AI prototyping, Figma-to-Code, Code-to-Figma and new handoff workflows.

30%+

estimated delivery-time reduction in selected initiatives.

An estimate, not an absolute metric. The bigger gain was not raw speed: it was less ambiguity between Product, Design and Engineering.

Available for opportunities

Open to senior roles and selected projects.

I am looking for teams that treat design as part of product decisions, and companies that need to turn complex problems into better digital products, experiences or operations.