Table of Contents
- Is AI changing UX design?
- What AI actually automates (and what it can't)
- You're not a producer anymore
- 3 important skills for UX designers in the AI age
- 1. Learn to train, not just use
- 2. Develop your editorial eye
- 3. Master the art of breaking things
- 3 types of UX designers
- Type 1: The producer
- Type 2: The integrator
- Type 3: The strategic director
- What junior UX designers need to do right now
- 1. Start using AI yesterday
- 2. Learn markdown
- 3. Stop waiting for perfect tutorials
- The AI tools worth knowing about
- The design industry is transforming
Do not index
Do not index
Read time: under 8 minutes
Is AI changing UX design?
UX designers had watched the job description get rewritten past few years. Not because they're getting replaced. Because the game just changed.
Chris just spent an hour with Sheamus Scott Grubb — Microsoft's Principal Design Director for GenAI, former design lead at LEGO and Sony, and here's what hit different: AI isn't killing design, it's killing busywork.
The question isn't whether AI will change your role. It already has. The question is: are you going to be on the right side of that change?
👉 The insight in this blog traces back to a thoughtful conversation with Shamus — here’s the full discussion:
What AI actually automates (and what it can't)
Most of what we call “design work” is actually production work.
Dark mode variations. Responsive breakpoints. Component libraries. Style guides. These aren't creative challenges. They're execution tasks. Important? Sure. But creative? Not really.
Sheamus is already training Claude to convert his designs to dark mode automatically. Take a design. Feed it to AI. Get dark mode variants. Done.
“That is work that we used to spend a huge amount of time doing, but is that really work we want to be doing? Is that what we got into design to do?” — He explains.
No. We got into design to solve problems, to create experiences, to make things that matter.
AI can produce. But it can't conceive, it can execute. But it can't strategize, it can generate variations. But it can't determine which variation actually solves the problem.
That's where humans come in.
You're not a producer anymore
The shift is subtle but massive: from creator to curator.
“The model is not bringing the craft, the model is learning instructions from us.” — Sheamus emphasizes.
Think about it like managing a junior designer. You don't tell them exactly which pixels to move. You teach them principles. You guide their thinking. You refine their output.
That's what AI needs from you. Not micro-management. Strategic direction. Here's how this plays out in practice:
Before AI:
- Spend 40 hours creating 10 design variations
- Choose the best one
- Refine it over another 20 hours
With AI:
- Spend 4 hours teaching AI your style
- Generate 100 variations in 30 minutes
- Spend 20 hours curating and refining the best options
Same quality. Different workflow. More time for thinking. Less time for executing.
3 important skills for UX designers in the AI age
The skills that separate good designers from great ones are completely different now.
1. Learn to train, not just use
Most designers treat AI like a magic button. Type prompt. Get result. Complain when it's wrong.
Wrong approach.
The right approach is to think of AI like an intern. You don't just tell them what to do once. You teach them your process, your standards, your style.
“The biggest skill that we can learn as designers is how to train these AI models to do what we want them to do.” — Sheamus says.
Sheamus built a custom tool for his team at Microsoft that chains together multiple AIs, taking article titles, generating metaphors, creating illustration concepts. Three pages of prompts. Hundreds of rules about what not to do.
“You have to have it double check its work three times before it publishes it, most of the bulk of the prompt was telling it what not to do rather than what to do.” — He notes.
That's the level of specificity needed. Not “make it pretty.” Not “design something cool.” Strategic instruction that teaches the AI your craft.
2. Develop your editorial eye
When AI generates 100 options in 30 minutes, your job isn't production anymore. It's curation.
Sheamus calls it being a “crate digger”, music producers who spend hours going through thousands of vinyl records looking for that perfect two-second sample.
“That's another mental model that we can start to use as a skill set. How do we look at this mass of material that this thing produces and find what is valuable within that?” — He explains.
Your new superpower: knowing what's good, knowing what solves the problem, knowing what connects with users.
Production skills become commoditized. Editorial judgment becomes premium.
3. Master the art of breaking things
AI wants to be perfect. Annoyingly perfect. Boringly perfect.
Sheamus spent months learning how to make AI produce “accidents”, the beautiful mistakes that make design feel human.
“AI hates making accidents. One of the things that I've worked on is how do I write prompts that force accidents like that?” — He says.
His discovery? Be less specific about subjects. More specific about style.
Want abstraction? Don't describe your subject in detail. “Woman on beach” works better than “a woman with long brown hair standing on a sandy beach with waves crashing behind her.”
“The more specific you are in your subject material, the less likely it is to abstract it. The AI just wants to prove that it can generate that temple in Kyoto that you really want to see.” — Sheamus notes.
Counter-intuitive. But that's exactly the point. The designers who win are the ones who figure out how to bend these tools away from their defaults.
Join 13,053+ Designers for FREE weekly UX Insights
Every Wednesday, I send out 1 actionable framework to grow your UX career 🌱 — No fluff. Always 2 minutes or less.
3 types of UX designers
This conversation reminded me of something Sheamus said early on:
“There are sprinters and marathon runners. And one of the things that I noticed is that agency people are really effective sprinters… I wanted to go in-house because I started to realize that a lot of my creative power was unlocked by the longer I spent with a specific brand or product.”
With AI, that divide gets more pronounced. Three types emerge:
Type 1: The producer
Still manually cranking out variations. Still treating Figma like it's 2015. Still spending 40 hours on tasks AI could do in 40 minutes.
These designers are in trouble. Not because they're bad. Because they're competing with tools that work 24/7 and don't need coffee breaks.
Type 2: The integrator
Uses AI tools. Generates stuff quickly. But treats it like a copy-paste machine. No refinement. No curation. Generic output with generic results.
Better than Type 1. But not sustainable. Anyone can use these tools. Using them well is what matters.
Type 3: The strategic director
Understands that AI is a tool, not a replacement. Focuses on the parts only humans can do: strategy, judgment, taste. Uses AI to handle production. Spends time on conception.
This is where you want to be. This is where the money is. This is where the interesting work lives.
What junior UX designers need to do right now
If you're early in your career and freaking out about AI, here's your action plan:
1. Start using AI yesterday
Load your portfolio into ChatGPT. Ask it how you compare to other designers. Ask what you could improve. Ask how to position yourself better for recruiters.
Take your resume. Feed it to Claude. Ask what keywords you're missing. Ask what story you're not telling. Ask what would make a hiring manager actually read it.
Not complicated. Just do it.
“The best piece of advice I can give to anyone is use the tools. Just start using it. The more you use these tools, the more you spend time with them, the more you'll start to see what they're good at, what they're weak at.” — Sheamus says.
2. Learn markdown
Want to write advanced prompts? Learn markdown.
“When you get beyond just like a one or two phrase if you really want these tools to follow your guidelines very strictly learn how to write them in markdown. That's actually how they write their responses to themselves.” — Sheamus explains.
It's not sexy. But it's effective. Markdown is how AI thinks. Learning to speak its language is like learning to code, except faster and more immediately useful.
3. Stop waiting for perfect tutorials
You won't find a comprehensive “AI for UX designers” course that teaches you everything. This space moves too fast.
“We're at a moment where no one really knows what works. We're in a completely blue space and we have to kind of understand that we're going to make a lot of beer drinking apps and fart apps along the way until we figure out what the pattern is.” — Sheamus notes.
Experiment. Break things. Figure it out as you go. The designers who succeed are the ones who start messy and iterate, not the ones who wait for clarity.
👉 A Designer’s Guide to Rule AI:
The AI tools worth knowing about
Forget the hype. Here are the tools Sheamus actually uses:
For prompting:
- Claude (his current favorite for document work)
- ChatGPT (still relevant, just not the only option)
- Use multiple AIs simultaneously: what works in Claude might not work in ChatGPT
For image generation:
- Flux models from Black Forest Labs (open source)
- Midjourney (with style references for consistency)
The design industry is transforming
Sheamus made one more point that stuck with me. We were talking about the transition from digital design to AI-powered design, and he brought up painting and photography:
“If you look at paintings before the invention of cameras and you look at paintings after the invention of cameras, they changed. For so long, the use of a painting was to replicate what the real world had. And I think most people thought paintings were going to die when cameras came out, but no. What paintings then started to move into is what is the emotion that you're trying to cover? What are the colors? What is the impression?”
Full abstraction came out because of the camera. Not despite it.
AI isn't killing design. It's forcing design to evolve into what it should have been all along: strategic, thoughtful, human-centered work that focuses on problems worth solving, not just tasks worth completing.
The question isn't whether you'll adapt. It's whether you'll adapt fast enough to be on the right side of this shift.
Your move!
Whenever you're ready, there are 4 ways I can help you:
1. Junior Designer Bundle: Transition into UX and craft an unforgettable portfolio & get hired.
2. Senior Designer Bundle: Become a design leader with systems to build healthier, happier teams.
3. UX Portfolio Critique: In less than 48 hours, get your 30-minute personalised video of brutally honest feedback.
4. Job Sprint Course: Stand out in an unpredictable job market by building a memorable personal brand and a killer job search strategy.
