Arrive with an idea. Leave with a validated, prototyped, AI-specced case study. Five live sessions across three weekends — research, prototype, test, implement, and defend it, all on your own product.
Limited to 20 founding participants · 3 weekends, August 29–September 12 · Recordings included · Future cohort price €399
Reservations are currently handled personally by email.
Prompt tricks age in weeks. What lasts is knowing where AI genuinely accelerates UX work — and where it quietly invents users that don't exist. This masterclass follows the UX process end to end: ideate, research, prototype, test, implement, defend. Every technique is demonstrated live on Second-Hand Classroom — a peer skill-exchange app ("trade what you know, learn what you don't") designed from concept to spec on research with real students — and then you immediately apply it to your own product idea.
By the finale you're not holding a certificate and a folder of notes. You're holding a validated concept, an insight report, a clickable prototype, and a completed AI Feature Spec — a mini case study for your portfolio, critiqued live by the curator.
Behind every screen there's a person, and they should feel happy — not confused, blamed, or interrogated — when using what we build. AI doesn't change that conviction. It just changes how fast we can honor it — or violate it.
One product — yours — carried through the entire arc. Each 90-minute session ends with a real artifact, and by September 12 the artifacts add up to a case study.
We open with an AI-assisted ideation block — sharpening your project brief, stress-testing the concept, and mapping its assumptions — so your fuzzy idea becomes a testable one in the first hour. Then the research foundations: interview guides, AI-assisted desk research, synthesizing raw notes into themes, and drafting personas from real data — with the hallucination traps flagged before you fall into them.
From Saturday's synthesis to a clickable flow in one day. Prompt → wireframe → prototype, Figma AI workflows, AI-generated copy and content design — and the judgment layer nobody teaches: when to trust the output, and when it's confidently lying to you. Your take-home for the gap week: polish the prototype and recruit two or three testers.
Your prototype meets real users. We plan a usability test around your riskiest assumption, then run the lab as peer testing — you test their prototype, they test yours — generating authentic session recordings and notes on the spot. AI analyzes the transcripts and survey data; you do the part it can't: turning messy findings into prioritized recommendations a team can act on.
The full anatomy of a shipped AI feature, demonstrated on the demo app: the Personalization Ladder, the prompt as a design spec, guardrails and GDPR, and the unit economics that decide whether the feature survives contact with a budget — including the question every 2026 spec must answer: what does your product know that free frontier AI doesn't? You fill in the same AI Feature Spec for your own product, section by section, folding in Saturday's test findings.
The finale. You present your AI Feature Spec plus one artifact from the earlier sessions, applied to your own product — and it gets critiqued the museum way: failures reviewed as exhibits, not verdicts. Memorable, specific, and never personal. This is where the case study earns the word "defended."
Your welcome email includes a short five-question brief — your product idea, its target user, and one problem it solves. It takes ten minutes and means Session 01 starts with your idea instead of a blank page. Arriving without one? No problem — you'll work on Second-Hand Classroom, the shared demo product, so nobody stalls.
Every session follows the same rhythm: watch the technique work on a real product, question it, apply it to your own project, then take it into your workflow. You leave each week with an artifact, not just notes.
The technique demonstrated live on Second-Hand Classroom — real prompts, real outputs, including the ones that go wrong and why.
Challenge the method. Bring the "but in my product…" questions before you start building.
Guided hands-on lab — you apply the same technique to your own project, step by step, with AI at your side.
Troubleshooting the lab — unstick anyone who got stuck, compare outputs and approaches across the cohort.
Folding the method into your real workflow — your team, your tools, your sprint. Plus the week's take-home task.
Two sessions per weekend, with a week between each block to do the work — after weekend one you'll polish your prototype and recruit testers; after weekend two you'll assemble the case study for the finale.
11:00 CEST · 10:00 London · 12:00 Athens. All sessions run live at this time and are recorded.
Mention your preferred time in the reservation email. If enough participants are US-based, a second live group opens at 11:00 ET — confirmed one week before we start.
Be part of the first live edition of the AI × UX Masterclass. Complete the full programme at a one-time founding rate and help shape future cohorts through your feedback.
Everything you need to keep the workflow running after the cohort ends.
90 minutes each, live on Zoom, with Q&A. Can't attend one? Every session is recorded and yours to keep.
The same structured spec used to design Second-Hand Classroom — personalization, prompts, guardrails, GDPR, unit economics — yours to reuse on every future project.
Reusable on every projectPresent your spec and artifacts in Session 05 and get structured, exhibit-style curator feedback.
Capped at 20 seats so every question gets answered and every critique gets time.
Research synthesis, insight report, prototype, spec — five artifacts that add up to a complete mini case study of your own product.
A Fragments United certificate for your portfolio and LinkedIn profile.
You're being asked to "use AI" without anyone defining how. Leave with a research-to-spec workflow you can defend — and a case study that proves you can run it.
You have an idea and no research team. Validate the concept, prototype the core flow, and spec the AI feature — before you spend a euro on development.
You can build anything AI suggests — the question is whether you should. Learn to test assumptions and spot AI's confident lies before they reach production.
Subscribe to the Fragments United newsletter and get the first Museum of UX volume instantly — 21 exhibits built from real interface decisions, recurring mistakes, and practical design lessons. It's the same critical eye the masterclass turns on AI-generated work.

I'm a UX researcher and consultant with 20+ years of experience helping teams create digital products users genuinely enjoy. I teach Human–Computer Interaction at Metropolitan University Belgrade — where I help students ideate, conceptualize, and carry their own product ideas from first sketch to defended work.
This masterclass runs the same way: you bring the idea, I bring the process — and twenty years of knowing where the process breaks. Because behind every screen there's a person, and they should feel happy using what we build — no matter how much of it AI helped create.
Five live sessions across three weekends: August 29 & 30 (Sessions 01–02), September 5 & 6 (Sessions 03–04), and the finale on September 12 (Session 05, live critique). The Europe track runs at 11:00 Belgrade time (CEST); a US-friendly group at 11:00 ET opens if enough enrolled participants prefer it — mention your preference in your reservation email.
Ideally, yes — that's the whole point. Before Session 01 you'll fill a short five-question project brief (your product idea, target user, one problem it solves), and every session's lab applies the technique to it. Don't have one yet? Session 01 opens with AI-assisted ideation to sharpen whatever you bring, and if you arrive empty-handed you'll work on Second-Hand Classroom, the shared demo product — nobody stalls.
Mention the US track in your reservation email. If enough participants are US-based, a second live group runs at 11:00 ET / 17:00 CEST for the same dates. Either way, every session is recorded, and the Q&A can be continued asynchronously.
Every session is recorded and shared within 24 hours, and you keep access to the recordings. The live critique in Session 05 works best live, though — plan for that one if you can.
Both, honestly — but differently. Beginners get the full UX process, compressed into three weekends with AI removing the busywork. Experienced designers get a defensible AI workflow — knowing exactly where the tools accelerate research, prototyping, and specification, and where they fabricate — plus a critique of their own work.
General-purpose AI assistants (ChatGPT, Claude, or similar) and Figma's AI features cover everything we do. Free tiers are enough to follow along and complete every artifact; if you already pay for a tool, you'll simply move faster. The methods are deliberately tool-agnostic — they'll survive the next model release.
It helps but isn't required. Volume I is the self-guided collection of UX exhibits — the same critical eye the masterclass applies to AI-assisted work, and the lens used in the Session 05 critique.
Full refund up to 7 days before the first session, no questions asked. After the first session, a 50% refund within 48 hours if it's not for you.
Validated, prototyped, AI-specced, and defended in a live critique — built on your own product across three weekends. Twenty seats, then the doors close until the next cohort.
Join the founding cohort — €99