Txoko Tabulo is the board-game café app I designed during my master’s. I took it from a generic app to one with personality, gamified and easier to understand, rebuilding the brand and every screen with AI connected to Figma in one evening. It only worked because the research was already done.

By hand, this redesign would have taken me 5 to 8 weeks. With AI and good research behind it, it took one evening.
The original app came out of a master’s project, with real interviews, focus groups and a usability test behind it.
AI didn’t do the thinking. The research already told me what wasn’t working and what people valued. Without it, there would have been nothing to redesign from, just faster screens.




Before opening Figma, I went back to the findings and turned each one into a design decision.
Tournament and open-table icons confused people in the usability test.
A colour and a game-piece shape for each section, plus a label on every card.
Interviews showed that what brings people back is the community.
A brand built on one idea: there's always room at the table.
The research asked for reasons to keep coming back.
Points, achievements with medals and a members ranking.
The AI read the original prototype and wrote straight into the Figma file: real variables, components and screens. I set the direction and reviewed every detail. The new brand keeps the cup people knew and adds the game they come for.
BeforeA cup with a face, bold colours and game-piece shapes.
Points, achievements with medals and a members ranking.
One colour per section and a label on every card.








It keeps the cup people knew and adds the game they come for.
Every change goes back to something users told us.
A full clickable prototype to put in front of regulars.
Research compounds. Interviews from my master’s became the brief. Good research keeps paying off years later.
AI multiplies judgement. It was fast because I knew what “right” looked like. Without that, I’d just have had more screens.
Next: test it. Five to eight regulars and the same five tasks as the first test, to check the problems are really gone.