How I Design in the Age of AI

✦ How I Turn Ambiguity Into Design Decisions

Most projects don't start with clarity — they start with scattered ideas, incomplete requirements, conflicting opinions, and a lot of unanswered questions.

Over time, I've built a process for moving from that uncertainty to a confident decision, and AI has become part of it — not to make the decisions for me, but to help me think faster, explore more directions, and get to a shippable answer sooner.

Here's how it shows up across my workflow.

✦ Making Sense of the Chaos

Claude · ChatGPT · Stakeholder Research

Every project starts with the same goal: getting to clarity before I open Figma.

I use Claude and ChatGPT to break down PRDs, meeting notes, and stakeholder feedback, surface the gaps, and identify the questions nobody's answered yet — feeding that synthesis back into research and stakeholder workshops rather than skipping straight to screens.

✦ Exploring Before Committing

Figma Make

Once the problem is clear, I use Figma Make to turn ideas into wireframes and flows in minutes instead of hours.

The biggest advantage isn't speed — it's flexibility. When a new direction costs ten minutes instead of an afternoon, I'm willing to explore more of them before committing to one, which is exactly how tasteX went from concept to a deployed UAT build in under 5 months.

✦ Refining the Decision, Not Just the Screen

Claude · Figma AI

This is where most of the real thinking happens.

I use Claude as a design-thinking partner — pressure-testing assumptions, checking flows against established heuristics, and arguing the other side of a decision before I commit to it. Figma AI's auto-layout and variant generation handle the repetitive structural work so I can stay focused on the product decision underneath it.

✦ Shipping Faster, Without Skipping Steps

Codex · Claude Code · GitHub Copilot

On projects where I'm also driving implementation, Codex and GitHub Copilot accelerate code generation, and Claude Code handles deployment and CI/CD — directing an AI-orchestrated pipeline rather than working through it screen by screen.

It's the difference between a three-portal product taking a traditional year to build and one that shipped in under five months.

✦ What AI Has Actually Changed

The biggest change isn't that I design faster — it's that I get stuck less.

I spend less time stalled at the start of a problem and more time refining the parts that actually matter, and I can explore more directions before I have to commit to one.

The tools will keep changing. The advantage is staying curious enough to keep changing with them.