A/B testing used to mean picking two options, splitting your audience, and waiting weeks for results. AI just changed the rules — now you can test a thousand variants simultaneously while you sleep. But faster testing comes with a hidden danger: AI will optimize for the wrong thing if you let it.

In this episode, Jill breaks down how multi-armed bandits, AI-generated variants, and real-time significance testing are transforming how travel brands — from cruise lines to theme parks — design better experiences. Plus, she shares a three-question framework you can paste directly into AI before any test to make sure you're optimizing for what actually matters.

Because AI runs the test. You ask the question. Don't mix those up.

• What A/B testing actually is (and why people overcomplicate it)
• 3 ways AI has upgraded the testing process
• Real travel industry examples: cruise bookings, airline fees, theme park notifications
• Why a higher click-through rate can be a trap
• A copy-paste AI prompt for better test design in 30 seconds

Zero budget. Timeline of yesterday. And a stakeholder asking if you can just ask ChatGPT instead.

Here's the thing — they're not entirely wrong.

In this episode, Jill walks through six free UX research methods, theme park edition, and shows how AI has quietly made most of them more powerful than ever. From mining hundreds of app reviews in ten minutes to guerrilla interviews in a theme park queue, these are real methods that work without a recruiting budget, a research lab, or a six-week timeline.

Covered in this episode:

— Review mining (and how AI does the affinity mapping for you)

— Guerrilla interviews with AI transcription

— The hallway test: why strangers at your company are your best testers

— Prototype testing when AI builds the prototype in two hours

— Observation journals — the one method AI genuinely can't touch

— Competitive analysis with AI-generated comparison matrices

Plus: Jill shares a ready-to-paste AI prompt at the end that matches any research question to the right method and spits out a 3-step execution plan. Instantly.

Next episode: Jill couldn't make it to the theme park, so she did the research from her couch instead. AI made it weirdly effective.

Design and Travel drops every Tuesday. Subscribe wherever you listen.

Most UX personas have a dirty secret: nobody actually talked to them.

Meet Sarah — 34, lives in Austin, loves hiking, frustrated by confusing checkout flows. You've made her. Jill's made her. We've all made her. And now AI can generate ten more Sarahs in under a minute.

But here's the problem: AI personas sound real. They have quotes, backstories, and emotional language — and that makes them more dangerous than a plain guess, because a guess knows it's a guess.

In this episode, Jill breaks down the UX NPC problem: why AI-generated personas are a double-edged sword, what happens when we stop talking to real people entirely, and the simple 3-Interview Gut Check that keeps your research grounded in reality.