UX South Africa

Two products can do exactly the same job, equally well, and one feels cold and templated while the other feels calm and considered.People are terrible at explaining that difference. They're extremely good at feeling it.In this talk, Jackie Zhang makes the case that beneath the visible layer of layout, type and copy sits an invisible one: the response all those decisions create. Reassurance while you wait. Confidence when money moves. Relief after a mistake.He walks through Uber's pickup screen, Apple Pay's four-step confirmation and Gmail's undo button, none of them magical on their own, all of them answering a question the user is already asking.Then he turns it on himself, taking apart roughly 245 attempts at his own portfolio and the moment he stopped asking what it should look like and started asking what it should feel like.He closes with a practical method: notice the reaction, name it, trace it back to a decision. Work out which moments actually carry weight. Then iterate for quantity, because the interesting ideas only show up after you've exhausted the obvious ones.And an argument for why this matters more now than it did two years ago: when AI makes everything easier to build, average becomes the default output. Memorable is what's left.Chapters00:00 Why Some Products Feel Alive01:31 The Uber Pickup: Designing for Uncertainty02:41 The Visible Layer and the Invisible One03:39 245 Attempts at a Portfolio04:04 The Japanese Book Cover Problem05:08 Asking What It Should Feel Like Instead05:23 Labelling the Qualities That Hold Your Attention06:15 What It Became07:52 How to Practise This: Notice, Name, Trace08:45 Where Should You Spend the Effort?09:51 Apple Pay: Confidence Through Clear Stages11:38 Gmail: Designing for Regret12:40 Iteration Is Where the Answers Come From13:00 Keep Familiar Things Familiar13:16 Why Quantity Beats Polish14:41 It's Not Delight Added at the End15:42 Q&A: Usability Gets Them In. What Brings Them Back?18:10 Q&A: The Fastest Way to Lose Trust19:50 Q&A: The Detail Nobody Notices Until It's Gone20:58 Q&A: Will Personality Be the Differentiator?22:50 Q&A: How He Built the Portfolio```ResourcesJackie Zhang on LinkedInhttps://www.linkedin.com/in/jackie-zhang-designer/Bashhttps://bash.com/UX South Africahttps://uxsouthafrica.com/UXSA Online Eventshttps://uxsouthafrica.com/events/uxsa-2026-july-onlineSubscribe for more talks from UX South Africa featuring leading voices in UX, product design, research, service design and AI.

Most bad products aren't born from stupidity. They're born from silence.Tomislav Kozačinski has spent over twenty years doing two things that turn out to be the same job: product design and stand-up comedy. On stage, feedback is instant - they laugh or they don't. In a meeting, nobody boos. They say "interesting," and six months later you've shipped something everyone privately knew was wrong.This talk is about the gap between those two rooms.Tom separates politeness from kindness and from respect. Kindness is fine, respect is mandatory. What he's after is the specific politeness that makes smart people nod along to bad ideas and turns conflict avoidance into roadmap.Through a run of stories, a feature built over two months for a client who never signed, a £30,000 annual invoice nobody had questioned, and the contract he lost for asking a CEO one honest question, he argues that the most valuable thing a designer does is ask what everyone in the room is too embarrassed to ask.One sentence that costs you nothing: I'm probably missing something, walk me through how this makes us money.And one rule that raises the bar: if your honesty doesn't end in a decision, it isn't honesty. It's commentary.Chapters00:00 Not Another AI Talk03:19 Bad Products Are Born From Silence04:09 The Language of Conflict Avoidance05:48 A Feature That Made Zero Money08:29 Sales Sold Something We Didn't Have11:07 The Sentence That Costs You Nothing12:56 What Stand-Up Comedy Teaches You13:43 "Interesting" Is Silence16:57 When Did You Last Remove a Feature?18:03 Build vs Buy: A £30,000 Question20:36 If It Doesn't End in a Decision, It's Commentary21:46 The Price: Getting Fired for a Question23:19 Agreeable Is a Commodity. Useful Is Scarce.23:53 The AI Workshop He Talked a Client Out Of26:10 The Invoice Always Arrives26:25 Q&A: Being Direct Without Being Rude28:58 Q&A: Standing Your Ground Without Making Enemies30:17 Q&A: One Behaviour to Remove From Product Teams31:25 Q&A: Honesty When You Can't Afford to Lose the Job34:29 ClosingResourcesTomislav Kozačinski on LinkedInhttps://www.linkedin.com/in/tomislav-kozacinski/Tomislav 's Consultancyhttps://www.kozacinski.com/UX South Africahttps://uxsouthafrica.com/UXSA Online Eventshttps://uxsouthafrica.com/events/uxsa-2026-july-onlineSubscribe for more talks from UX South Africa featuring leading voices in UX, product design, research, service design and AI.#UX #UXDesign #ProductDesign #DesignLeadership #ProductStrategy #DesignCritique #StakeholderManagement #UXSouthAfrica

Some of the most damaging AI failures aren't caused by bad actors or broken systems. They happen when a system works exactly as it was designed to.In this talk, Dr Lizette Spangenberg unpacks what Eric Meyer called inadvertent algorithmic cruelty, the harm that surfaces when an algorithm does its job perfectly and nobody stopped to ask who it might hurt.Drawing on examples from Spotify Wrapped, Facebook's Year in Review, targeted advertising and healthcare AI, she shows how harm enters at three different layers of a product, data, logic and interface, and why UX teams usually only get involved at the last one.She then walks through a real case study: a face-scan health app that simply could not detect a colleague's face because of her skin tone. And she asks whether a disclaimer at the bottom of the screen is ever good enough.Whether you're a designer, researcher, product owner or BA, you'll leave with a sharper set of questions for evaluating any AI product: who does this work worst for, what happens when it's wrong, and what are we optimising for, at whose expense.Chapters00:00 Unpacking the Title00:43 Algorithms, AI and Machine Learning04:04 Spotify Wrapped and the Pink Pilates Princess07:25 Targeted Ads After a Breakup08:43 When Your Photos App Remembers Too Much09:59 Year in Review and "Algorithmic Cruelty"12:24 They All Worked Exactly as Intended13:30 Where Harm Enters: Data, Logic, Interface14:46 The UX Gap in AI Projects15:23 Three Questions Every Team Should Ask17:07 The Black Mirror Test17:55 Case Study: Face Scan Health Apps20:08 When the System Can't See Your Face22:28 Are Disclaimers Good Enough?24:07 Accuracy vs Equity25:17 What Designers Are Responsible For27:09 The Ring Ad: When AI Sees Too Much28:57 Q&A: Where Does Responsibility Lie?31:48 Q&A: Is Anyone Getting Guardrails Right?34:15 Q&A: Biased Data and the Human GuardrailResourcesDr. Lizette Spangenberg on LinkedInhttps://www.linkedin.com/in/lizette-spangenberg/X, bigly labshttps://dischemgroup.com/xbigly/UX South Africahttps://uxsouthafrica.com/UXSA Online Eventshttps://uxsouthafrica.com/events/uxsa-2026-july-onlineSubscribe for more talks from UX South Africa featuring leading voices in UX, product design, research, service design and AI.

A yoga instructor in a village hall in Northamptonshire drew out every pose on a sheet of paper so that one person in her class could follow along. Twelve years later, Joanna Pawluk still remembers it.In this talk, Joanna traces a line from that piece of paper to the work she does today at Emirates, where she's mapping the end-to-end travel experience for customers who are deaf or hard of hearing.Along the way she makes a point most organisations miss: the people who experience the most friction become the best at hiding it. They build their own workarounds, they get from A to B, and the service looks like it's working fine.Drawing on her lived experience of hearing loss, desk research, social listening and workshops with frontline crew, Joanna shares how she built a service blueprint two metres long, and how she turned it into a reusable framework any team can apply.That framework flips the usual question. Instead of mapping every customer type, you ask one thing of every touchpoint: if a sense or an ability were removed here, would someone still get the same quality of service?Chapters00:00 Introduction: A Personal Story01:01 Cape Town, 2009: Teaching With a Globe02:32 Designing Something They'd Remember03:33 Living Between the Hearing and Deaf Worlds05:21 A Yoga Class in a Village Hall06:09 Rosa's Piece of Paper07:29 She Followed a Design Process Without Knowing It09:42 Travel: The Hardest Service Design Challenge10:51 People Who Experience Friction Hide It Best11:45 How Do You Know When to Board?13:31 Starting the Project at Emirates14:50 The Research Behind It17:42 Why Social Listening Tools Miss the Visual19:33 A Spectrum of Need, Not a Persona21:29 A Service Blueprint Two Metres Long23:49 A Reusable Framework for Any Team25:31 Closing: Stop Making People Adapt26:43 Q&A: Where Organisations Overlook Accessibility28:10 Q&A: Which Industries Lead and Which Lag29:35 Q&A: The Biggest Opportunities AheadResourcesJoanna Pawluk on LinkedInhttps://www.linkedin.com/in/joannampawluk/Emirateshttps://www.emirates.com/UX South Africahttps://uxsouthafrica.com/UXSA Online Eventshttps://uxsouthafrica.com/events/uxsa-2026-july-onlineSubscribe for more talks from UX South Africa featuring leading voices in UX, product design, research, service design and AI.

Everyone in the room agreed. They thanked you for the work. The research was solid. And a few weeks later, nothing had changed.If that's familiar, this talk is for you.Across 11 years in healthcare, aviation, enterprise systems and consulting, Ignatius Ferreira watched excellent UX practitioners struggle to create impact while others moved things forward with ease. For years he put it down to design maturity, or stakeholders who just didn't get UX.Eventually he realised he'd been treating two very different skills as if they were one: understanding a problem, and creating the conditions for that problem to be solved.In this talk, Ignatius argues that we've become exceptionally intentional about understanding users, and almost entirely unintentional about understanding the organisations our work has to survive inside. He unpacks how to learn an organisation the way we learn a user — how trust forms, how priorities compete, and why some ideas move while others quietly disappear.Three lessons anchor the session: stop communicating like a UX designer, treat relationships as research rather than influence, and when someone pushes back, ask what they're trying to protect.Chapters00:00 A Provocative Title01:24 The Model I Believed In02:24 Two Completely Different Skills03:28 The Assumption Hiding in the Model04:07 Everyone Agreed. Then Nothing Happened.05:07 Blaming the Organisation06:09 Nobody Was Trying to Stop Good UX07:34 What Was Missing From My Old Talk08:24 How Organisations Actually Behave09:46 Organisations Are Human Systems Too10:51 Day One: The Invisible Organisation11:48 Every Conversation Is a Piece of the Puzzle14:31 The Room Fills Up and You See a System15:28 Being Right Isn't Enough16:22 "What Would Have to Be True?"17:06 Agreement Is the Starting Line17:35 Clients Buy Confidence, Not Deliverables19:24 Lesson 1: Stop Communicating Like a UX Designer20:48 Lesson 2: Relationships Are How You Learn an Organisation21:53 Lesson 3: What Are They Trying to Protect?23:04 Closing: One Question24:24 Q&A: Building Influence Without Authority26:58 Q&A: South Africa vs the Middle East27:58 Q&A: Politics vs Healthy Organisational Dynamics30:28 Q&A: The Turning Point32:34 Q&A: Which Skill Mattered MostResourcesIgnatius Ferreira on LinkedIn⁠https://www.linkedin.com/in/ignatius-ferreira/⁠Emirates⁠https://www.emirates.com/⁠UX South Africa⁠https://uxsouthafrica.com/⁠UXSA Online Events⁠https://uxsouthafrica.com/events/uxsa-2026-july-online⁠