Hey and welcome back to a new week!

In this issue:

🏗️ INTERESTED IN DESIGN SYSTEMS & AI? JOIN ME!

Together with Level Up Club and Femke.design I am running a session on the AI workflow I tested and established at my own work to turn a messy Figma-only design system into a robust pipeline of components.

I’ll touch on technical, design-specific but also organizational topics on how I facilitated this and how you can do so too in your own organization.

This session is hosted by the wonderful Julia Fernandez (Meta).

When? August 26, 6PM (CET) / 12PM (EST) / 9AM (PST)

How much? Just $10!

“AI-native” is quickly becoming one of those terms people put everywhere without explaining what it means.

You see it in LinkedIn headlines and portfolio introductions. You see people calling themselves “AI Product Designers” because they use ChatGPT occasionally. I would be careful with that.

If you genuinely design AI products or have worked deeply on AI-assisted features, say so. Be specific and keep it fairly low-key. But using AI tools does not automatically make you an AI product designer, and calling yourself AI-native does not prove that AI has meaningfully changed your work.

The term itself is a little strange. It seems to borrow from “digital native,” which described generations that grew up with computers and the internet as an ordinary part of life. Most of us did not grow up with generative AI. We wrote, researched, collaborated, and made decisions without it. Even many students entered university before these tools became part of everyday work.

I do not want to split hairs over the definition. The bigger problem is that the label is hollow.

If your portfolio says you are AI-native, my next question is simple: Where can I see that?

We touched on this in the article about how hiring managers scan portfolios. A claim in your hero needs evidence underneath it. So reverse the order: do not start by branding yourself as AI-native. Start by doing work that makes the conclusion obvious.

There are roughly three layers through which you can demonstrate that.

The first layer is using a general-purpose LLM as part of your normal design process. That could be ChatGPT, Claude, Gemini, Grok, or something else. The model matters much less than people think.

You might use it to work through research data, explore ways to frame a problem, identify patterns in feedback, brainstorm edge cases, structure a workshop, or improve interface copy.

This is quickly becoming table stakes. It is useful, but simply writing “I used Claude” will not make a case study impressive. You also do not need to mention AI in every section. Mention it where it naturally affected the work.

For example:

That tells me you used AI to accelerate a heavy task without outsourcing your judgment. You could make the result more visible by turning the research into an interactive artifact or visual summary that lets people explore the themes.

The important part is the combination of speed and responsibility. LLMs can miss nuance or invent plausible connections, so verify the output against the source. Be careful with confidential company or participant data too. Using AI does not remove your responsibility for the result.

If you want to gain some extra points here make use of things like Artifacts in Claude and mention / show them. Especially in the research scenario above this would be a very nice use of AI by showing the research results visualized.

The second layer is the most valuable one for your portfolio and on the market in general. It includes AI features inside design tools such as Figma or Paper, plus tools that turn ideas into working software: Lovable, Codex, Claude Code, Cursor, and whatever comes next.

The visible difference can be enormous. One designer says, “I built a prototype,” and shows a Figma file with dozens of noodles connecting slightly different frames. Another shows a responsive, coded prototype with real inputs, useful states, validation, and interactions that feel like a product.

The second example is stronger—not because every project needs code, but because the designer used the available tools to make the idea more tangible and testable.

We recently covered how to turn a Figma design into a live prototype, so I will not repeat the workflow. For your portfolio, focus on making the capability visible:

You can return to an older project too. Pick its most interesting interaction, build a working version, and add it as a later exploration. Just be honest about the timeline:

A playground section is another strong option. Build a small app, interaction study, visual experiment, or useful little tool. It does not need a 5,000-word case study. It needs to be focused, polished, and interesting enough that someone wants to try it.

Connect the experiment to the designer you want to become. If you care about interaction design, build an interaction. If you care about data-heavy products, make a small data tool. If you care about AI products, prototype an AI-assisted workflow instead of adding another chatbot to a mobile app.

And please apply your design judgment. AI can generate generic interfaces and an impressive quantity of mediocre work. Do not publish the first result. Edit it, simplify it, test it, and make it yours.

The third layer is more advanced. This is where AI does not only help with an isolated task; it operates as part of a broader workflow, sometimes continuing work with less direct input from you.

You might have agents that compare a coded product with its design system, detect token drift, test components, update documentation, or prepare implementation work for human review. A self-healing design system is one possible direction: a system that notices when design and code drift apart and helps propose or apply a fix.

That is exciting. It is also not where I expect most junior designers to be.

If you have not built a live prototype yet, do not pressure yourself to create an autonomous multi-agent workflow. Roughly 80% of the practical value right now sits in Layer 2: using AI to design, prototype, build, and make better work visible.

As a junior, awareness of Layer 3 can be enough. A design-system case study might end with a reflection like this:

That can start an interesting interview conversation. But only use terminology you can explain. If “agentic,” “autonomous,” and “self-healing” are decorations you cannot discuss, leave them out.

And honestly, if you have already built a robust system that improves your product while you sleep, you probably do not need this guide. You might need to write the next one for me.

You do not need to prove that you use every new tool. You need to show that you recognise where AI improves the work—and where it does not.

Before adding an AI claim, ask:

If the answer is yes, include it. If the only evidence is a list of model names in your About section, leave the label out.

To start this week, pick one strong project and find one place where AI could make the work more tangible. Turn an important flow into a coded prototype. Create an interactive artifact from the research. Build a small companion tool. Keep the scope small enough to polish, then explain what the tool enabled and what you learned.

That is already more convincing than writing “AI-native” in a giant font.

AI fluency will matter more and more in design. But the people who stand out will not necessarily be those who talk about AI the most. They will be the ones whose work shows that they can use it with purpose.

Use general LLMs to accelerate the basic work while keeping ownership of the thinking. Use design and coding tools to make ideas real, testable, and more ambitious. Learn about agentic workflows, but do not rush toward complexity before the earlier layers feel natural.

Let the evidence come before the identity.

The strongest proof that you understand AI is not a label in your hero section. It is work that would have been slower, narrower, or impossible without it—and that still feels designed by someone with judgment.

🗣️ I ALMOST DO EVERYTHING BY YAPPING THESE DAYS

ChatGPT is a superpower if you know how to use it correctly.

Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.

Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.

Download the free guide

Cindy Ly’s portfolio feels playful without losing focus.

She is currently studying HCI and Design at the University of Washington and has already completed a product design internship at Zendesk. The work fits comfortably alongside some of the strongest student portfolios I have seen from that program.

The overall concept leans into an operating-system aesthetic. That direction has become more common over the last few years, but Cindy gives it enough of her own treatment to keep it from feeling copied.

The taskbar, navigation patterns, subtle control surfaces, and small interface details create a recognizable system without overwhelming the work. That balance matters.

A portfolio should feel personal, but it still needs to work for the people reviewing it. Cindy manages to do both.

That’s it for this week—thanks so much for the support! ♥️

Do you want your own portfolio reviewed in-depth with a 30-minute advice-packed video review? Or do you require mentoring to figure out a proper strategy for your job search?

I got you!

Keep kicking doors open and see you next week!
- Florian