How I workfrom first question to working prototype
The tools I use and what each one is for, how a project takes shape, and the principles behind my decisions.
Tools, and what each one is for
The tools I reach for first, and what each one is for.
Primary tools
Claude
Core workflowPerplexity
ResearchFigma
Design + prototyping
Lovable
Functional buildsGoogle Stitch
Early UXGPT/Gemini
VisualsThe rest of my stack
Lovable
Adobe PhotoshopHow a project takes shape
The stages that show up on every project, in order.
Research & context
How I approach this depends entirely on the project. One project called for tearing apart competitors screen by screen, figuring out what held up and what was worth borrowing. Another meant going straight to the audience with a survey to understand who would use the product and what they needed.
Concept to final design
Concepts take shape first in Nano Banana, Google Stitch, ChatGPT or Lovable, depending on what the task calls for. Then the work moves into Figma with AI still in the loop: Figma Make and Figma Agent for generating directly in the file, and Figma MCP when Claude needs to work on it.
Image & visual content
Midjourney, ChatGPT Images and Nano Banana get used interchangeably for references, concepts and finished visuals that go straight onto the site. Canva comes in for quick formatting when a visual needs adapting. If a project needs an image, I make it.
How I think about my work
If a project needs an image, I make it. Stock libraries are a last resort.
I don't run the same research playbook on every project. What I need to know changes case by case, so the method does too.
When I'm the only designer on a project, I also own the brand voice and how it comes across.
I generate several different directions with AI early on, then pull the best of each into one strong version.
I switch tools whenever something new solves the problem faster or better.
I like to walk through the actual method behind a piece of work, including the steps that led to the result.
