Before designing, I need to understand which problem we’re solving.
Before moving into digital product design, I worked for several years as a Civil Engineer. That experience shaped three ways of working: thinking about how a system connects, designing within real constraints and paying attention to how people use it.
Today I apply them to design: I organize information and flows, prioritize feasible solutions with engineering and consider the context of use and accessibility from the start.
How I design, stage by stage
I use the double diamond as a reference, not a recipe. What matters to me is separating two moments: understanding the problem well, and then exploring how to solve it.
Understand
I start by talking to users, reviewing data and understanding how the product works today. I also need to know the business and technology constraints before proposing a solution.
Claude helps me organize interviews, spot recurring themes and prepare syntheses. Then I review the findings and go back to the sources before turning them into decisions.
What to research, who to talk to and which signals really matter.
Focus
With enough information, I turn what I learned into a concrete problem. I prioritize what to solve now, what to leave out and how we will know if the solution works.
I use it to compare patterns, organize scenarios and explore different ways of structuring information.
The priority, the scope and what really goes into the first version.
Explore
I try different ways of solving the problem before committing to one. I work with flows, wireframes and prototypes until I find a direction that makes sense for users and the team.
I use Claude Design and ChatGPT Sites to test ideas quickly and Lookback Eureka AI to speed up the review of testing sessions.
Which ideas are worth testing, with whom, and when we have enough evidence to move forward.
Build
Once the direction is clear, I work with engineering to turn it into product: states, edge cases, responsive behavior and the details that appear when a screen stops being a prototype.
Claude Design and ChatGPT Sites help me build functional prototypes and websites closer to the real product.
What gets specified, what can be simplified and when a technical constraint means the design needs another look.
I use AI to move faster, not to let it decide for me.
I use it to organize research, explore alternatives, summarize information and build prototypes faster. But I don’t take an AI result as evidence: I go back to the data, check it with users and the team and decide what makes sense to keep.