Modern Skill Stack is a podcast by Uxcel CEO Gene Kamenez. Real stories on how digital product professionals are keeping up with a constantly changing skills landscape.
Listen on YouTube.
There’s a line Elizabete dropped about ten minutes into our conversation.
“You will not be replaced by AI as a designer. But designers who use AI might replace you.”
Easy to nod at. Easy to forget. What I kept circling back to was the mechanics, what it actually looks like when you’re using AI well, where most designers go wrong, and why Elizabete thinks the entire conversation about tools and prompts is already missing the point.
Elizabete founded Collectif Studio, a consultancy built around helping companies move from ambiguity to execution, specifically in the AI transition. Before that, over a decade as a UX and service designer. Startups. Large organizations. A stretch at Accenture Song, where design decisions carry real weight and the margin for messy thinking is thin. Her career moved through the UK, Germany, the UAE, and a meaningful period in Denmark, where she developed real grounding in co-design and facilitation. Those skills are now the core of how she approaches work and teaching.
She’s now a regular workshop facilitator at Uxcel. Her hands-on AI workshops sell out quickly, and consistently received 4+ rating. This is why I wanted her to walk me through what she actually teaches, why that structure, and what she wishes more designers understood right now.
This episode is brought to you by Uxcel.
Uxcel is a learning platform used by 500,000+ product professionals and 200+ organizations to build cross-functional skills across UX design, product management, and AI. Learn through short, gamified courses that fit into a working day, earn certifications, and build a portfolio with real-world project briefs.
Start learning at uxcel.com
You’ll learn:
Why AI works best as an intern and fails badly as a director, and what that means for your design workflow
How one team cut a 20-hour copywriting task down to 2 hours without sacrificing quality
The “20% upfront” rule Elizabete uses before touching any AI tool
Why mindset matters more than knowing the right prompts
The workshop structure she uses to help designers overcome fear and build real AI habits, and the structure you can adapt for your own practice
Timestamps
00:00 Intro
01:17 Importance of strategic AI orchestration
03:38 Elizabete’s career journey
06:01 Transition from working 9-5 to starting a studio
07:12 How to use AI to reduce boring work
10:59 Spend the first 20% of AI time doing this
13:01 The importance of quality input
15:01 Live workshop walkthrough
22:43 Overcoming blockers in AI workflows
26:33 Final tips for designers using AI
27:29 Outro
My biggest takeaways
AI is a great intern and a terrible director
The framing that cracked this open for me: Elizabete doesn’t think about AI as a threat or a replacement. She thinks about it the way she’d think about a capable junior: someone who can execute a lot, but who absolutely should not be left in charge of decisions.
In Elizabete’s words:
“AI is a great intern and a terrible director. The decision making always stays with me or with the decision makers in different companies. It’s really about utilizing AI to give those boring tasks that you don’t want to do.”
Hand AI a vague prompt and expect it to figure out what you actually want, and you’re setting yourself up for mediocre output and real frustration. Treat it like an intern who needs a clear task, a defined scope, and honest feedback, and it becomes something you can actually rely on.
Try this next time: before you open any AI tool, spend a few minutes writing down what you actually want the output to look like. Not “help me with my research.” Something closer to: “I need three distinct user personas for a 30–45 year old audience who are lapsed gym members. Each persona should include a core motivation, a primary frustration, and a quote in first-person voice.” That shift in how you brief the tool changes everything you get back.
The real ROI isn’t time saved, it’s where that time goes
Elizabete shared a specific case: a team whose designers were spending around 20 hours on copywriting tasks. Necessary work, but not the kind that requires a UX professional’s judgment.
In Elizabete’s words:
“I helped the team speed up their process from 20 hours to two hours on copywriting. Give the task to AI, provide the right background information, and the team can spend their time on more valuable and strategic tasks.”
The number is definitely striking. But the more interesting part is what she says next, where the time goes. Getting 18 hours back means nothing if those hours just fill with other low-value work. The question worth asking first (before you touch any tool) is: what are the high-value activities your team is currently skipping because the tedious stuff is taking up most of the calendar?
The designers who thrive are the ones who were already solving problems
When I asked Elizabete whether job anxiety partly comes from losing the comfort of busy work, knowing exactly what you’re supposed to do each day, she didn’t soften it.
In Elizabete’s words:
“The skills that will be valued in the age of AI are really this critical thinking and problem solving. If you’ve always been a designer who solves problems, this is an amazing time for you because you can give all these more boring tasks to the AI. But if for the last year or so, or if you’re maybe a new designer coming in and you’re really focused more on the pixel pushing, then it’s the right time now to invest more into these problem-solving skills.”
This tracks with what I see on our side at Uxcel. The designers who feel most grounded about their futures aren’t the ones with the longest tool list. They’re the ones who can frame a problem clearly, who know what a good solution looks like before they’ve built anything, who can make judgment calls when the AI output is close but not quite there.
If you’ve built those skills, AI mostly just gives you more hours. If you haven’t, it makes the gap more visible.
Spend 20% of your AI time before you start
The sessions where AI frustrated Elizabete most were the early ones, before she worked out what was actually going wrong.
In Elizabete’s words:
“I learned how critical it is to spend the first 20% of your working time with AI to define what you want to get out of it and for what purpose. Whenever I work with a tool, it’s important to first define the right prompt, to really understand what I want to get out of it.”
Twenty percent sounds slow when you’re trying to move fast. But on a two-hour task, that’s 24 minutes of thinking before you type a single word into the tool. What does a good output actually look like? What context does the AI need that it won’t have by default? What would a bad output look like, so you can catch it?
Most people do the opposite. Open the tool, type something approximate, get something approximate back, spend an hour trying to salvage it. The 20% rule front-loads the thinking, and it’s counterintuitively faster.
To be honest, I’m still not sure I’ve fully adopted this habit. I know the rule but I find myself skipping it when I’m in a rush.
Mindset is the skill, tools are just the vehicle
The part of the workshop structure I found most interesting wasn’t which tools Elizabete chose or which prompts she favors. It was what she said the workshops are actually teaching.
She runs a two-task structure: an individual research exercise followed by a prototyping task, outputs from the first feeding directly into the second. Participants pick from three project clusters, a dating app, a food delivery product, a fitness app, enough creative distance from their day job to actually experiment without everything feeling high-stakes. After each task, the group shares what they built, how they prompted, what surprised them.
But when I asked what the workshop is really about:
In Elizabete’s words:
“This isn’t really about the prompts and the tools. Of course those matter, but it’s more about the mindset. Understanding what you can do with AI, and focusing on practical ways to improve your workflows instead of being blinded by the AI hype.”
I’ve seen enough designers running in place to know this is right. The hype creates a kind of paralysis: you feel like you should be using these tools, but you’re unsure which one, or for what, or whether the one you chose is already obsolete. The shift Elizabete is after is simpler than that: stop hunting for the perfect tool. Find one task in your current workflow that AI could lift off your plate. Do that one thing. Learn from it. Then do the next.
That’s how habit formation actually works. More durable than any prompt library.
“Garbage in, garbage out” — and why foundations matter more now, not less
There’s a version of the AI conversation I keep running into: now that AI can generate UI, draft copy, and synthesize research, strong foundations matter less. Just learn the right prompts.
Elizabete pushed back. So did I.
“It’s the garbage in, garbage out principle. If your input is garbage, the output will be garbage.”
If you don’t know what a well-structured persona looks like, you can’t tell when AI hands you a bad one. If you don’t understand what makes a prototype testable, you can’t brief the tool to build something useful. If you can’t recognize a strong research insight, AI-assisted synthesis just scales your confusion faster.
The designers getting the most out of these tools know enough to give good direction, and enough to evaluate what comes back. That’s a fundamentals story as much as an AI one. Maybe more.
A note from Gene
Every episode of this podcast starts with the same question: what’s the one skill you’re personally focused on right now? Elizabete’s answer was “strategic AI orchestration”. Not tools, not prompts, but learning how to look at a design workflow and understand where AI belongs and where it doesn’t.
That’s the same question Uxcel was built to help product professionals work through.
We’ve built courses specifically on AI skills for product teams, structured to fit into a real working day. Five-minute lessons, practical exercises, real certifications you can point to.
Start learning at uxcel.com
And if you haven’t subscribed yet, Modern Skill Stack drops new episodes with practitioners working through these questions in real time. Subscribe on YouTube.
Resources and tools mentioned
Workshops:
AI as a UX Co-Designer (Uxcel community workshop by Elizabete): https://luma.com/UxcelAIUXCoDesigner
AI + Personas deep dive (upcoming Elizabete workshop): https://luma.com/UxcelAIUXPersonas
Where to find Elizabete:
Collectif Studio: https://www.collectif.team/
Uxcel workshops: https://luma.com/uxcel
Production and marketing by Uxcel. Interested in appearing on the Modern Skill Stack podcast? Reach out at partnerships@uxcel.com
What’s the one task in your current workflow you’d hand off to AI first, and what’s stopping you? Share it in the comments.
Cheers,
Gene
