The first time we asked ChatGPT to help with a UX interview guide, it felt like talking to a clever intern who’d skimmed Wikipedia and thought they understood the field. The response was generic, awkwardly formal, and just off. It didn’t understand nuance. It didn’t ask questions back. It couldn’t clarify its assumptions. And honestly? That made sense. Early AI models were built more like autocomplete engines.

But today’s models are different. They can reflect tone, infer context, follow chains of reasoning, and stay coherent across multi-step instructions. We’re at the beginning of something seismic. Something we haven’t quite found the language for. And everyone is talking about becoming AI-first. What does it mean to be AI-first—not just as a business or a brand—but as an academic, a designer, a knowledge worker?

Don’t miss out on our livestreams and weekly posts, hit subscribe and be part of a growing community of people, who love UX and AI.

To be AI-first isn’t just about learning tools. It’s about unlearning old habits. For decades, academics were trained to polish prose, cite sources, and agonize over whether that paragraph needed a semicolon. But now? Now we’re entering an era where the most powerful form of scholarship might be a well-written prompt. The keyboard is no longer just a writing tool for us, but it’s a command line to an infinite and never-tiring assistant that can automate away the parts of our job that we hate or at the very least find a little too tedious.

Being AI-first means approaching your own thinking like code. It means treating ideas not as fixed endpoints, but as modular, remixable inputs. It means asking, constantly: What parts of my job are the thinking bits—and what parts are formatting, summarizing, synthesizing, transcribing, or presenting? And if a machine can do the latter 90%, should I still be spending most of my time there? Probably not.

Academia was built on the industrial model of knowledge: slow, peer-reviewed, incremental. But AI moves at the speed of questions. And if we don’t start asking better ones, we’ll be left behind.

Being AI-first doesn’t mean outsourcing your intelligence. It means upgrading your interface to it. It’s a design problem. A workflow challenge. A shift in epistemology.

And once I started seeing prompting that way—as an act of design, not delegation—it changed everything for me. And in our video livestream that we hosted this week, we shared 10 simple but powerful rules for writing better AI prompts—especially if you’re a UX designer or researcher. We called them the “10 Commandments of AI Prompting.” (Yes, cheesy. But effective.)

Let’s walk through them together. With examples.

1. Be specific

Asking AI to "write a marketing plan" is like asking someone to cook dinner without telling them if you're vegan or allergic to peanuts.

A better prompt: “Write a 1-page marketing plan for a note-taking app targeted at college students, focused on social media and referral campaigns.”

Specificity reduces irrelevant output. More signal, less noise.

2. Break big tasks into bite-sized pieces

AI can handle a lot but it still loves structure.

Instead of: “Analyze these transcripts.”

Try: “Step 1: Summarize the key themes from these user interview transcripts.

Step 2: Identify three recurring pain points.

Step 3: Propose design directions based on those pain points.”

It’s like writing a to-do list for your AI. Everyone wins.

3. Set the scene with context

Context helps AI know where it’s going and where you’re coming from.

Don’t just say: “Make this landing page better.”

Say: “I’m designing a landing page for a productivity app aimed at ADHD professionals. The goal is to increase trial signups. Here’s the current copy: ”

Context gives the AI necessary information to create better output.

4. Ditch ambiguous language

Words like “improve,” “optimize,” or “better” are fuzzy. Be concrete.

Instead of: “Improve this UX copy.”

Say: “Make this onboarding screen more concise and friendly for first-time users.”

Precision is clarity. Clarity is power.

5. Define your goal

If you want great answers, say what success looks like.

Don’t just ask: “Generate survey questions.” Ask: “Generate 8 open-ended survey questions that help me understand why users abandon the signup process halfway through.”

When AI knows your destination, it picks a better route.

6. Respect AI’s limits

Pick the right AI for the job. Claude, ChatGPT, Gemini, Grok—they each have strengths.

Claude is great for creative writing and code. ChatGPT nails structure. Gemini 2.5 Pro has a massive context window and is currently the overall best model. Use the right tool.

Also: Don’t be afraid to try the same prompt across different AIs. Compare and steal the best bits. Remix your output.

7. Write for a real audience

AI is great at role-playing. Use that. Constraints force creativity.

Try: “Act as a UX researcher. Review this usability test plan and suggest improvements.”

This nudges the AI into the right direction where your audience is. You can also tell it to write for a specific audience. Clear limits = tighter results.

8. Show examples

AI predicts best from examples. Give it a model to mimic.

Want better headlines? Include three examples of your brand’s tone. Want tighter summaries? Show what “good” looks like.

AI is a pattern matcher. Give it a pattern worth matching.

9. Embrace the feedback loop

Iterate and iterate. Your first prompt won’t be perfect. That’s fine.

Treat it like prototyping: try something, see what works, tweak, repeat.

Even changing a single word (“summarize” vs. “analyze”) can shift the whole result.

10. Keep it simple, keep it clear

Big words don’t make you sound smarter. They make your prompt harder to interpret.

Try: “Design a rewards system that’s simple, motivating, and doesn’t rely on points.”

That’s it. No jargon. No corporate nonsense. Just plain, instructional human talk.

Final Thoughts

Prompting is all about designing questions, providing steps to a solution, and doing this all with the right clarity so it is easy to understand.

And just like in UX, the better the design of your prompts—the better the experience.

So next time you’re staring down a blank AI chat box, remember our 10 commandments. To part the sea and all that good stuff. Let us know which commandment you already follow.

BTW: Are you watching other designers land $150K+ roles while you're stuck doing the same old UX process? ​The truth is brutal: In 2025, there are only two types of UX designers...

* ​Those who've mastered AI and charge premium rates.

* ​And those who haven't and compete on price.

Don't be number 2. For the first time ever, we’re revealing the exact $150K AI-UX Blueprint we’ve kept behind university walls in the last 2 years. Sign up here to join live for this Masterclass on June 11. Only 20 spots.