UX Digest - best posts

I Thought I Already Knew How to Do Research A reflection on the difference between collecting findings and actually synthesizing them into insights. The key shift is building a traceable chain from evidence to patterns, interpretations, and product decisions NNG: Do Participants Say Less to AI Than to Human Moderators? In a controlled study, participants spoke 45% less with AI moderators, while AI-led interviews were 37% shorter and much more moderator-dominated. The result suggests AI moderation can reduce participant elaboration, although more talking does not automatically mean better insights Prototyping: We’re Trying Hard to Design the Right Solution, But Did We Design the Right Path? A product can be built for the right audience while the path to its value still ignores their needs. The author uses Inflow’s 37-step onboarding for people with ADHD to show why reducing friction and demonstrating value before collecting sensitive data can matter AI: Vibe coding alone Vibe coding let the author turn a decade-old budgeting spreadsheet into a working tool without a developer — and made her question what gets lost when everyone starts replacing colleagues with AI. The punchline comes later: the tool wipes her data, and Claude can’t help her recover it Opinion: Why Researchers Need to Build Researchers lose context when insights pass through reports, designers, and developers before becoming a product. Learning to prototype and build lets them turn research directly into testable solutions without becoming full-time engineers Design: If You Don’t Get Lost in a City, a Designer Probably Did Their Job Well Copenhagen’s public transport shows how wayfinding, consistency, and progressive disclosure can make a complex system feel effortless. Good design gives people the right information at the right moment and often works best when users barely notice it Basics: Low Fidelity - A Concept First Approach to Design Low-fidelity prototypes help teams test the concept before visual polish makes a weak idea look convincing. Their value is not speed itself, but exposing gaps, inviting critique, and making problems cheaper to fix early @uxdigest

What Are UXers’ Ways of Understanding Product Features Four lenses for evaluating any feature: the job it does, the behaviour it shapes, its place in the system, and the friction it creates. The best question may be the simplest: if this feature didn’t exist, what would users actually do instead? The Problem with “UX Theater” Research becomes UX theater when it produces workshops and reports but never changes the product. A sharp rule from the article: share the decision, evidence, and next step — not a 100-page deck no one will open The Devaluation of Discovery As execution gets faster and cheaper, teams are increasingly tempted to skip discovery and jump straight to solutions. The uncomfortable consequence: AI can help you build the wrong thing faster than ever 🎟️ NNG: Neurodivergence - What is it and Why Does it Matter Neurodivergent users can face challenges with executive function, communication, and sensory processing. The practical takeaway: design for simplicity, clarity, control, and support — because neuroinclusive design often improves the experience for everyone Tools: 10 UX Research Tools for Different Research Methods in 2026 A practical map of 10 research tools matched to specific methods — from recruitment and moderated testing to diary studies, IA, analytics, and repositories. The core rule is simple: choose the research method first, then pick the smallest tool that supports it without duplicating your stack AI: 8 voice AI UX patterns A map of 8 emerging Voice AI patterns — from avatar calls and AI phone agents to dictation and ambient assistants. The useful part is how each pattern changes the agent’s perceived role, and with it the UX rules for turn-taking, interruption, memory, and recovery Prototyping: The Death Of The Button - Why The Best Interface Is No Interface Intent-driven design replaces chains of menus, filters, and forms with one goal the system executes for you. The provocative shift: future UX may be judged less by interaction quality and more by how completely the interface disappears @uxdigest

Persistent lessons in human-centered automation A 1980s photocopier study and a grocery-store robot point to the same lesson: people interact with automation based on what they think it can do, not what it was designed to do. In the agentic era, managing expectations and making system intent visible may matter more than adding another interface NNG: Test Complex Interactions Earlier with AI Prototyping AI prototyping makes it possible to test complex interactions — filters, dashboards, conversational AI — much earlier with realistic working prototypes. In one Ramp case, the team found a confusing edit-tracking issue that static screens simply couldn’t reveal Prototyping: Coupons, Vouchers, Gift Cards, and Credit - Four Things That Look Alike and Aren’t Coupons, vouchers, gift cards, and credit may look interchangeable in an interface, but they follow different rules and user expectations. Treating them as one pattern can create surprisingly confusing checkout experiences AI: We Used to Test Buttons. Now We Test Chaos AI makes UX testing less deterministic: the same input can produce different outputs, errors, and paths. So instead of checking one “correct” flow, researchers increasingly need to test ranges of behaviour, edge cases, and how gracefully the system fails Case Study: Bypassing Institutional Distrust - UX Lessons from Redesigning Philanthropy A case study on designing philanthropy around distrust rather than trying to “fix” it with better UI. The strongest insight: money may flow upward to institutions, while trust flows sideways through local communities — so the service had to follow the trust, not the org chart @uxdigest

The UX of Anxiety Reduction Good UX can reduce “micro-anxiety” by making actions predictable, reversible, and easier to understand. A useful shift: design not for attention, but for intention — so users know what will happen next and whether they can recover from a mistake Effectiveness of AI Moderators: A Literature Review A review of 10 studies finds AI moderators roughly comparable to humans on procedural tasks, but weaker where judgment matters — especially probing and knowing when to go off-script. The verdict is nuanced: AI looks promising for structured, high-volume research, but not yet for discovery and emotional nuance NNG: AI Can Help Write an Article, but It Can’t Stand Behind It NN/g uses AI for rewriting, formatting, repurposing content, and even critiquing arguments — but not for final judgment. Their core rule is simple: AI can help think, but a human expert still has to understand, evaluate, and stand behind every published claim Process: How to Structure a Jira Backlog — Epics, Stories and Priorities That Actually Work A practical guide to keeping Jira useful instead of turning it into a graveyard of tickets: clear hierarchy, ruthless ranking, and detailed work only near the top. The sharpest rule: polishing a story you won’t touch for six months is usually waste because it will change before you get there AI: From the Red Dust to the Agent Age A researcher reflects on moving from ethnography in rural Africa to automating parts of research with AI agents. The key lesson: before automating research, you first have to make expert judgment explicit — including rules for what counts as an insight and when contradictions must go back to a human Case Study: Designing RootCause to Make Plant Diagnosis Less of a Guessing Game A case study of RootCause, an AI app that diagnoses sick plants from a photo and turns confusing online advice into clear care steps. A nice detail: a plant chatbot was added not just for answers, but to make diagnosis feel less stressful and more human @uxdigest

UX in 2027 may be less about interfaces and more about behavior A look at how UX may shift from designing fixed flows to defining how adaptive and agentic systems behave. The article covers generative interfaces, AI-assisted research, design systems for machines, trust, permissions, reversibility, and accessibility. The central idea is worth keeping: designers may draw fewer predetermined paths and spend more time defining the boundaries within which software is allowed to create its own NNG: We Still Need Long-Form Copy Short attention spans haven’t killed long-form content: some topics are simply too complex to compress without losing clarity or trust. NN/g argues that longer copy still matters when users need depth, confidence, or strong SEO and AEO signals. The useful part is not “write more”, but structure long content so people can scan it through summaries, bullets, callouts, bolding, and visuals AI: How people looked for information using AI A usability study with 20 disabled people and family members on how they actually search with AI, Google, and other sources. People switched tools depending on the task: AI was more attractive for complex, personalised questions, while trusted websites and search remained important for high-stakes topics. One especially telling finding: some participants blamed themselves when AI gave a poor answer, assuming they had simply written the wrong prompt Writing: Billing approvals that don’t feel like a trap - the exact warning microcopy patterns Billing microcopy should answer five things before the user confirms: how much they’ll pay, when they’ll be charged, what changes, whether they can undo it, and whether there are extra costs. The article breaks this into practical patterns like “you’re about to”, “here’s what changes”, and a receipt preview. One especially useful detail: showing the exact path to downgrade makes reversibility feel real instead of theoretical Basics: The UX of Human Connection - Why Client Experience Starts Before the Product UX starts before anyone sees the product — in the first meeting, the tone of communication, and the questions a team chooses to ask. The article argues that curiosity and informal human moments can uncover better insights than polished presentations or rigid agendas. In one CRM project, a simple question about how salespeople actually start their day helped the team avoid building several unnecessary features @uxdigest

Nine Design Fixes for Your UX Research Reports Nine practical ways to make research reports easier to read and harder to ignore: clearer structure, stronger hierarchy, better contrast, more whitespace, consistent charts, and reusable templates. The authors make a very UX-ish point — a research report is itself an interface with two users: the researcher creating it and the stakeholder reading it. MeasuringU even usability-tested its own report template and went through four iterations before settling on the current version The user belongs to everyone, which is the same as belonging to no one When everyone on a product team “owns the user”, responsibility for actually understanding them can disappear. The author shows how overlapping research by UX and Product can turn into a data-quality problem when interview framing and methods aren’t shared. His practical fix is surprisingly small: agree upfront who researches, who is consulted, and who makes the call NNG: Using AI for UX Work - Study Guide NN/g’s practical map of where AI already fits into UX work — from research planning and analysis to prototyping, writing, service design, and workshops. The guide is less about “AI will replace UX” and more about choosing specific tasks where it can genuinely help without outsourcing judgment. The research section alone covers synthetic users, AI-moderated interviews, survey writing, digital twins, and why analysis still shouldn’t be handed over completely to AI Experience: Learnings from Experience Design Micro-Projects at NID A two-week experiment in designing tiny experiences around emotions like restlessness, adventure, and optimism — often with just a day to research, prototype, and test. The projects range from an endless staircase timer to chocolate as a trigger for hope and motion-activated “fireflies”. The best lesson comes at the end: even a handful of people was enough to break assumptions and reshape almost every prototype AI: The feature is cheap to code, so we might as well build it. (Probably not) AI makes features dramatically cheaper to build, but implementation is only a fraction of their real cost. Every new feature adds maintenance, support, friction, complexity, and eventually makes removing it a product decision of its own. The uncomfortable takeaway: when almost anything can be built, deciding what not to build becomes the harder skill Opinion: What first dates can teach us about becoming better, safer researchers What first dates can teach researchers about rapport, awkward silences, knowing when to end a conversation, and trusting your instincts. The authors connect familiar dating situations with practical research skills — from making interviews feel less like a checklist to planning fieldwork with exits, check-ins, and even a code word. A useful reminder that researcher safety deserves as much preparation as participant safety Basics: Persona Is Not a Document, It’s a Decision Tool A persona is useful only when it helps a team make product decisions, not when it simply describes age, job title, and favourite apps. The author suggests building personas around context, goals, pain points, behaviour, motivation, and decision drivers — then using them to prioritise features and UI choices. A good test is simple: can your persona help explain what you should avoid building? @uxdigest

Unlocking UX Secrets: How 1 Hour of Session Recordings Transformed Trendyol Seller Panel Just one hour of watching session recordings (Hotjar) revealed a massive pain point: sellers suffering the "scroll of death" for bulk actions, forced to scroll back to the top after selecting items. Three iterative designs later, a bottom-positioned sticky toolbar triggered by checkbox selection won—proving that quick, unbiased observation in natural workflow can uncover insights no survey or interview could, without spending extra budget or bothering busy users NNG: The Custodial Era of UX - Cleaning Up After AI AI lets teams build faster than UX can evaluate, creating UX debt: interfaces that "technically function" but confuse users, overload features, and skip foundational questions about whether something should exist at all. UX's new custodial role isn't just janitorial cleanup—it's triaging what to keep, accelerating evaluation (rapid user panels, checklists), and adding UX knowledge to generation (Design.md, design systems)—because the goal is to prevent the next mess, not just clean it up Prototyping: Rethinking Data Visualisation - A UX Approach To Dashboards That Actually Drives Decisions Most dashboards fail because they're built from available data, not from a specific operational question—so they show numbers without driving decisions. The fix is upstream: define context (what question are we answering?), audience (familiarity + accountability), and insight (what should change after seeing this?), then design the visualization to serve that, not the other way around—because the right chart isn't about simplicity, it's about appropriate complexity for the person and the decision at hand AI: Thoughtful AI implementation for UXR leaders AI skeptic or not, set a north star: AI should support, not replace, research quality (craft, assessment, revision). Preserve core skills with clear guidelines (don't use AI for research questions, do use it for data cleaning, label AI output), and frame every tool conversation around risk vs. reward—because in the rush for efficiency, we risk losing the human processing time that produces real insights Design: My “Book of Sorcery” Taught Me Design Before the Internet Did Before the internet, a fine art diploma taught design through observation and limitation: hand-drawn ads, lettering from a book of 30 fonts, and a physical "book of sorcery" of clippings from magazines and packaging. That instinct to collect and trust what "looks right" later met research training—behavioral psychology, HCI, accessibility—transforming instinct from the answer into the first question: why does it look right, to whom, and under what assumptions? @uxdigest

The gap between what user research surfaces and what gets built It isn't a quality problem—it's a translation problem: qualitative insights don't become backlog items because there's no formal process to bridge the formats. The fix is to embed decision-makers in synthesis, translate findings into prioritization-ready format (impact + effort), and track their status like bugs—otherwise, research remains an expensive activity with no measurable value NNG: AI Can't Replace Real Research in Empathy Mapping AI can help organize existing research into empathy maps, but it cannot replace real user evidence—because plausible, generic quotes (e.g., "I wish the app would ask before swapping") are not the same as specific, messy details from real sessions (e.g., "It swapped my oat milk for whole milk and charged me before I saw the notification"). Before reaching for AI, ask: am I organizing real data or generating data I don't have? If the answer is the latter, the map needs more research, not more AI Prototyping: Using content design to fix UK emergency alerts The UK emergency alert about wildfires was well-intentioned but failed its users: dense text, redundant language, weak visual hierarchy, and unhelpful signposting made it harder to scan in a moment of stress. A content redesign—shorter sentences, direct "do not" commands, clear paragraphs, and a specific URL—cut reading time from 19 to 14 seconds and dropped the reading age from 17.6 to 14.6, proving that in an emergency, every word and second counts Experience: We need a survey — a UX research story When a stakeholder says "we need a survey by Tuesday," the real skill isn't asking for more time—it's starting with the question, not the method, and doing credible research within imperfect conditions. Fast research means reducing scope, focusing on the most critical questions, and making deliberate trade-offs (e.g., 5 relevant users instead of 20 later)—and sometimes the method itself evolves as findings emerge, because research is about following the evidence, not sticking to a plan AI: Craft still matters, but it’s about outcomes Craft isn't about pixels anymore—when AI makes production cheap, real craft moves to choosing the right problem and curating AI output with taste. To keep it, write your standards down (so the machine applies your rules), keep research a weekly habit, and always put a human review at the end—because AI fails quietly, and fluent confidence without verification is just fiction Design: Why People Ignore Perfectly Designed Buttons People ignore perfectly designed buttons not because of color, size, or placement, but because of the trust problem (uncertainty about what happens next), decision fatigue (too many choices already made), and identity questions (not ready to be the person who clicks it). The fix isn't polishing the button—it's building the right emotional state, trust, and narrative momentum before it appears, because a button is just the answer to a question the experience needs to ask first @uxdigest

Five Ways to Use Rapport When Moderating Good moderation isn't just reading a script—it's building trust through preparation (know your audience's lingo), deliberate ice-breaking, and a verbal toolkit (purposeful pauses, neutral expressions, camera eye contact). But watch for the failure mode: mutual people-pleasing that makes disagreement feel impolite, and use rapport as a final screening step—because a few warm-up questions can catch a participant who never should have qualified Why Users Don’t Experience Your Interface the Way You Designed It The gap between your Figma file and reality is shaped by the curse of knowledge (you can't unsee your design decisions), context collapse (users are distracted, stressed, rushing), mental model mismatches (they expect different logic), and accessibility oversights. The fix isn't more testing, but treating it as genuine inquiry: test in real contexts, involve people who disagree with you, and cultivate humility—because users experience your design through their own reality, not yours NNG: We Still Need Long-Form Copy Despite shrinking attention spans, long-form copy still matters—because some things are inherently complex, require trust, or need SEO and AEO benefits. The key isn't cutting words, but making them scannable with summaries, bullet points, callouts, bolding, and visuals to improve comprehension and engagement with content over 1,000 words Prototyping: Designing High-Performance User Experiences for Data-Heavy Dashboards and Analytics Products Consumer design patterns (clean cards, progressive disclosure) fail in data-heavy expert tools—experts need density, bulk actions, and state preservation, not simplicity. The craft lies in balancing density, clarity, and speed: dense tables with strict alignment, keyboard shortcuts, and transparent permissions—because friction isn't clicks, but the gap between how work actually happens and how software forces it AI: Researcher-in-the-loop Instead of asking if AI will replace researchers, ask who keeps research honest when everyone can self-serve. The researcher-in-the-loop model makes researchers governors of the system: they set rules (sourcing, confidence, escalation), route low-risk questions to AI tools, and keep high-risk ones for humans—because a confidently wrong AI is worse than no answer Design: How Good Design Will Save the World Good design isn't loud—it's invisible, seamless, and accessible, solving problems without being noticed. The best designers focus not on articulating every use case, but on delivering what people really want (like the snooze button: a simple reprieve in a complex world), and make accessibility the core, not an afterthought—because true design considers both the space and the user, creating experiences that work for everyone, not just the able-bodied @uxdigest