MCP lets you connect AI to the research repository you already have—the best of both worlds. Here's the bill it hides, and four questions to run before you connect.
Buy vs build with AI? Either way you're going to pay.
Claude f*cked with my head
Pride in the AI Pipeline: 10 Ways to Keep AI From Erasing Queer Users
Don't Frankenstein Your Flow: How to Integrate AI Into Research Workflows with Integrity & Intention
The Map Is Not the Terrain: What the AI-Native UXR Frontier Leaves Out
Jess Holbrook just published the clearest map of AI-native UX research I've read (Holbrook, 2026). If you lead a research team and you haven't read "Frontier UX Research Circa May 2026," go read it first, then come back.
He gets the big thing right: being AI-native is a systems problem, not a tools problem. Buying Dovetail or shipping one Claude project won't make you AI-native. You have to redesign how you work — how evidence flows from intake to insight, collection to synthesis to decision, so AI can participate at every step. I've been making the same argument to researchers and clients for a year, in nearly the same words: context engineering over prompting, systems over tools, amplification over automation. Human judgment and growth are the things you protect.
So this isn't a rebuttal. It's an extension. A "yes, and…" if you will.
