Briefs usually arrive half-solved.
Someone already decided the homepage needs a redesign, when the real problem is that nobody can explain the pricing. I spend the first weeks of a project checking whether the brief points at the right thing.
Services
Product & UX strategy
Figuring out what the product is, who it's for, and why the current version confuses people. Interviews, journey mapping, positioning. This happens before anyone opens Figma.
Digital product design
Flows, interfaces, and the boring-but-decisive details: what a form asks for, in what order, and what happens when someone gets it wrong.
Brand-led experience
Products where what the brand promises and what the interface does are the same thing. When they drift apart, users notice before the team does.
Design systems
Reusable patterns so the fifth product page doesn't get designed from scratch. Also so developers stop asking which gray this is.
Principles
Strategy before visuals
At dhosting I spent the first month on interviews and support tickets before touching a single screen. The screens went faster because of it.
Clarity is a conversion problem
People don't buy what they don't understand. Most of the “UX problems” I've been hired to fix turned out to be comprehension problems wearing a UI costume.
Show, don't hand off
I teach interaction design at the Academy of Fine Arts in Gdańsk. Explaining why a decision is right, to students or to stakeholders, is half the job.
Process
Find what's actually broken
I read support tickets, talk to sales, and interview users before forming an opinion. At dhosting the support team's ticket categories turned out to be a better research tool than my first round of interviews. I found that out later than I should have.
Make it testable fast
A coded prototype beats a debate. I build them with Claude Code, usually within a day or two of having something worth testing, so decisions get made on how people actually click, not on how the meeting went.
Keep brand, product, and UX in one conversation
These usually live in three teams that meet quarterly. The product suffers in the gaps between them. Part of my job is being the person who has read all three roadmaps.
Check whether it worked
Shipping is not the end of the project. I look at what changed: fewer tickets, better completion, faster time to first value. When nothing changed, that's information too.
How AI fits into this
This part changed my job more in two years than anything in the previous six.
My design system lives in Figma as tokens. Claude Code reads it through Figma MCP and builds against a CLAUDE.md file that encodes my spacing scale, typography, and implementation rules. This site was built that way, in Astro. There was no handoff document. There was nothing to hand off.
What that means in practice: an idea becomes a coded, clickable prototype in hours. Research synthesis that used to eat a full day of sticky notes takes an afternoon, and I spend the saved time on the part AI is bad at, which is deciding what matters.
I keep a running note of where AI actually saves me time versus where it just feels fast. So far the honest answer: big gains in prototyping and synthesis, almost none in framing the problem. Not once has a model looked at a brief and said the brief is the problem.