brb, getting you some chai :)

How I use AI in my practice

Day-to-day

Beyond the things I've built, AI is woven into how I work. Roughly by job:

Research and synthesis

Gemini for research. It's extensive and thorough. NotebookLM for working with interviews: on the Testing Toolkit, I put every user interview (voice, video, notes) into one NotebookLM, so I could reference any of it at any point in the project.

Thinking and copy

ChatGPT for quick, compact, to-the-point input, like refining a line of copy or getting a rough thought out. It tends to just agree with me, though, so for anything I want genuinely pushed on, I go to Claude.

Building

I use Claude and AI agents to build small tools that fix real workflow problems, and I'm comfortable working with GitHub through AI agents. When something small can save the team real time, I'll build an agent for it. The Customer Feedback Analyst is one.

UI exploration

Gemini's canvas for getting UI directions flowing quickly. I've used Figma Make too, it wasn't for me.

The throughline

I'm a deliberate user. I pick the tool for the job, I have opinions about where each one falls short, and I build with AI rather than just prompting it.