UXLx 2026 Wrap Up: Talks Day Notes

15 May was the final day of UXLx 2026, which means it's Talks Day! 🤩 Ten industry leaders took the stage to share ideas and insights with an engaged audience of UX professionals from 29 countries, all coming together to learn, connect, and recharge their creative energy.

From AI and leadership, to conflict navigation, inclusive collaboration, content and the future of UX, the talks explored some of the most pressing and exciting topics shaping UX today.

This wrap up is your chance to revisit the key takeaways if you joined us in Lisbon. If you couldn’t make it this year, you'll get a glimpse into the conversations driving the future of our industry.

💡 Read through the article or click on a talk title to jump straight to the session that interests you most.

🕙 MORNING

• Stepping Up to Leadership: How UX designers can grow their influence and advance their career - Doug Powell
• Mindful Conflict Navigation: A Service Designer's Guide - Sylvie Abookire
• Inclusive Collaboration Against Algorithmic Bubbles - Zalihata Ahamada
• Content Modelling for Scalable Reuse: Build Once, Publish Everywhere - Noz Urbina
• Ancient Wisdom for Modern Design: Technology changes fast. Human behavior doesn’t. - Sarah Thompson

🕝 AFTERNOON

• Designing for AI: UX Patterns, Practice, and Product Differentiation - Dave Brown
• Beyond Experience: Who Benefits From the Systems We Design? - Ayesha Moarif
• The Wobbly Chair - Brandon Harwood
• Designing What Matters: Beauty, Data and the Human Touch in the Age of AI - Marianne Ashton-Booth
• AI Stupefaction (And What We Can Do About It) - Christopher Noessel

Stepping Up to Leadership: How UX designers can grow their influence and advance their career by Doug Powell

Research Study Overview

• 4-month study by Washington University MDES (Master of Design for HCI and Emerging Technology) students on design leadership career patterns
• 135 survey responses + 12 in-depth interviews with design leaders globally
• Mapped 93 distinct career pathways from education through current leadership roles
• Goal: understand and map career journeys and identify emerging leadership archetypes

Key Insight 1: Tool Stack Disruption

• 93% expect their design tool stack to change completely or significantly in next 5 years
• Creates fundamental instability in design practice foundation
• Cascading effects on team composition, methods, metrics, and leadership approaches
• Opportunity: new practices haven’t been invented yet - room for everyone to be pioneers
• Design Leadership Core principle: “Change is the new normal” - adaptability essential for future design leaders

Key Insight 2: Mentorship Gap Crisis

• Only 21% of design leaders had strong early-career mentorship
• 41% had no early career mentorship at all
• Compare to other professions: 85% of new lawyers receive mentoring early in their careers, 75% of business executives say mentoring was critical to their success, 86% of CEOs credit mentoring as critical force in their career
• Common sentiment: “I was not prepared at all for my first leadership role”
• Opportunity: current leaders must intentionally build systems of support for next generation
• Design Leadership Core principle: Isolation is the worst enemy of design leaders. Combat isolation - design leaders need intentional relationship building with other designers, and less time with cross-functional peers and stakeholders

Key Insight 3: Burnout & Portfolio Careers

• 49% identify burnout as major career setback (alongside lack of growth, insufficient support, layoffs)
• 38% indicate career uncertainty - planning exits, transitions, or unsure about future
• Shift from linear career paths to diversified “portfolio” models
• Less than half plan to continue current trajectory over next 5 years
• New model: “portfolio of practices” - fractional roles, consulting, teaching, advisory positions
• Portfolio approach creates career resilience through diversified income streams and reduced employer dependency
• Design Leadership Core principle: “Empathy first with each other, then with our users” - need to turn design empathy inward

Positive Outlook & Next Steps

• Recent trend: design leaders “breaking the surface” after being underwater
• Building skills and confidence to move with current rather than fight against it
• Designer skills will be essential for navigating complicated months and years ahead - to create new models of leadership, new practices of support and new possibilities
• Full research study to be published end of summer

Sketchnote by Chris Noessel

Mindful Conflict Navigation: A Service Designer's Guide by Sylvie Abookire

Conflict Navigation Framework

• Conflict navigation vs management/mediation - goes deeper to explore underlying needs, identities, positions, worldviews
  • Management = referee approach (temporary, resumes when intervention stops)
  • Mediation = compromise approach (split the orange, often unsatisfying)
  • Navigation = explores root causes, creates accountability and new opportunities
• Designers uniquely positioned for conflict work - bridge diverse stakeholders, have empathetic creative methods
• Collusion concept (Arbinger Institute) - both parties perpetuate conflict by inviting/provoking what they’re fighting against
  • Creates positive feedback loops that escalate simple misunderstandings into hostility

Practical Application Tools

• Positions, Interests, Needs framework
  • Position = what people say they want (tip of iceberg)
  • Interests = the “why” behind the position
  • Needs = fundamental values, worldviews, emotional needs driving everything

• Designer mindsets for conflict
  • Question assumptions (same skill used for “are we solving the right problem?”)
  • Seek to understand (research/discovery applied to conflict)
  • Withhold judgment, replace negative interpretation with curiosity

• Disrupting collusion cycles
  • Notice narrative building and confirmation bias
  • Replace “they’re lazy/don’t care” with “that response was unexpected”
  • Ask questions to build relationship before trying to get needs met

Self-Awareness Exercises

• Mindfulness practice - tune into inner world before engaging conflict
• Reflection prompts: conflict superpowers vs vulnerabilities
• Recognition that conflict is both difficult/uncomfortable AND inevitable/universal
• Systems thinking to identify feedback loops in conflict dynamics

Sketchnote by Chris Noessel

Inclusive Collaboration Against Algorithmic Bubbles by Zalihata Ahamada

Understanding Organizational Bubbles

Four types of bubbles that shape workplace decisions:

• Informational bubbles - curated social media, AI newsletters, personalized feeds that reinforce existing thinking
• Organizational bubbles - different teams (design, tech, product, sales) have separate realities, rituals, information sources
• Design teams often seen as “startup within company” that others don’t understand
• Cultural bubbles - domain-specific norms for dress, language, presentation styles, documentation practices
• Developers write ADRs and document decisions; marketing/design teams have less written tradition
• Defensive bubbles - protection from conflict and suffering
• Quote from public admin project: “I prefer staying in my bubble because it keeps me from suffering”

Reframing Conflict and Decision-Making

Mary Parker Follett framework (1925) for handling conflict:

• Domination - one party imposes, but conflict returns as resistance
• Compromise - each party gives something up, nobody satisfied
• Integration - find real needs behind positions, create new solutions
• Window example: instead of opening halfway, find different rooms for different needs

Key reframing questions when decisions feel wrong:

• What context makes this make sense?
• What constraints are they seeing that I don’t see?
• What risk are they trying to prevent?

Practical Applications

Figma vs PowerPoint case study at previous company:

• Design team used Figma slides, executives wouldn’t read them or access Figma
• Solution: stakeholder needs mapping workshop covering design team, product owners, leadership
• Created PowerPoint templates with brand elements that looked like design team but used executive-friendly format
• Result: opened doors to business communities that didn’t know design team existed

AI considerations:

• Makes everyone sound more alike (conformity effect)
• Decreases accuracy on tasks outside AI capabilities, but people over-trust it
• Use AI to widen questions and lower contribution barriers, not to close debates or make decisions

Sketchnote by Chris Noessel

Content Modelling for Scalable Reuse: Build Once, Publish Everywhere by Noz Urbina

Omnichannel vs Multichannel Strategy

• Multichannel: Getting content out in various formats (website matches PDF matches app)
• Omnichannel: Channels work together additively to create bigger, better experience than individually
• User-centered approach vs publisher-centered
• Each channel used for what it does best in concert

AI and Content Quality Connection

• Content quality = next frontier of AI performance improvement (not just bigger models)
• Models need high-quality context to avoid hallucination and generic output
• Data/content is what makes AI unique to specific business rules and operations
• Training data was the internet, which is made of content

Component Content Approach

Core concept: Moving from document management to component management

• Also called: modular content, blocks, topics, atomic content
• Works like Lego - build bigger components, connect components
• Replaces document metaphor with reusable blocks
• Enables governance and standards for fields and UI strings

Implementation Examples

Omnichannel at scale (pharma company example):

Step 1 - Source material is split into 3 layers 

1. Content layer: Actual words + video statements for presentations
2. Data layer: Variables (age, region, references - EU vs US CDC)
3. Presentation layer: How it looks, completely separate

Step 2 - Staff select desired combinations of content, data, and presentation

Step 3 - System produces output 

• Single source → 16 outputs (4 formats × 4 regions)
• Staff wizard: select content pieces + countries + channels = automated generation
• No manual crafting of individual PDFs

Event management example:

• Problem: Events with multiple dates, each was a page, showing up multiple times on event lists (bad UX)
• Solution: Self-referencing parent/child model
• Parent passes down main info (speaker, overview)
• Children add specific details (venue, specialized CTAs)
• Enables automated mailing campaigns

Tools mentioned:

• Contentful (headless CMS) - content only, no presentation
• Figma plugin ($12) - pull content into design layouts
• WordPress with specialized content types

Scaling Results

• Customers generating 150+ outputs from single source
• Translation to 70+ languages
• Manage and approve content once, deploy everywhere

Designing Content Models

• Apply a design process to words
• Process:
• Diagnosis - market and user research
• Model - before writing, the requirements (how should components be built?)
• Ideate - content model, reuse strategy, personalisation strategy
• Draft - create content and designs
• Deliver - test 
• Measure - optimise and scale
• Domain model - the relationship and roles of various real-world concepts in a particular domain
• Content type - building the design of the LEGO pieces, the structural relationships and roles of various content types working together.

Sketchnote by Chris Noessel

Ancient Wisdom for Modern Design: Technology changes fast. Human behavior doesn’t. by Sarah Thompson

AI anxiety reality check:

• Endless stream of new tools (overwhelming as builders and as users)
• Over 1 million AI research papers published last year
• Endless courses
• Technology pace exponential vs. human brain evolution (hundreds of thousands of years)
• Gives stable foundation to design from - the human on other side of technology remains same

Behavioral Science in AI Era

• Core insight: Human brain hasn’t evolved in 40,000 years - same brain as cave dwellers, but now dealing with exponential AI advancement
  • All decisions are emotional (System 1 decides 10 seconds before conscious awareness)
  • 180+ cognitive biases affect users, but can be simplified into 6 factors
  • Companies applying behavioral economics see improvement in sales growth and gross margin (Gallop)

Emotional ROI Model

Designing for the emotional part of the brain

Framework: System 1 predicts costs vs gains across 6 factors - more gains = action, more costs = avoidance

1. Mental - Thinking is hard, brain burns most calories per ounce

• Costs: forms, choice overload, deciding what to prompt AI
• Gains: defaults (95% never change), Netflix “play something” button
• When you’re designing: Can the user understand what they need to do in seconds?

2. Social - Wired to belong, social pain = physical pain in brain

• Costs: leaderboards (for people at bottom it’s a social cost, feeling judged)
• Gains: AI chatbots for sensitive topics (less judgment than humans)
• When you’re designing: Does this make users feel safe, seen and part of the group?

3. Emotional - Automatic associations and reactions

• Use of colour (form errors, prices)
• Images- Process images 6-600x faster than text
• Gain: Duolingo example: big eyes, bright colours, doesn’t look like learning
• When you’re designing: What automatic reaction might this design trigger?

4. Physical - Abilities and senses;  Conserve calories, avoid effort

• Costs: small tap targets, eye strain, showing people exercising
• Gains: when you reduce physical cost, you see participation increase - tap to pay, facial recognition, passive data collection (Oura ring)
• When you’re designing: Can I remove real or perceived physical effort?

5. Material - Money, resources, tools; Scarcity mindset, accumulate resources

• Endowment effect: “we’ve gifted you 2GB data” vs “reply yes for free 2GB” (2x activation)
• Digital streaks act as currency
• When you’re designing: Are we asking users to give something up - or are we giving them something in return?

6. Temporal - Crave immediate rewards, hate waiting

• 100ms delay = substantial drop-off (Google experiment)
• Domino’s pizza tracker manages perception of time
• When you’re designing: Can I give an immediate reward or make people feel like they’re saving time?

Case Study: American Heart Association

• Problem: Heart Walk fundraising goals weren’t increasing despite UX improvements
• Issues identified: No emotional appeal, required typing, time consuming, asks to lead not follow, too many decisions, asked for money twice
• Solution: Heart slider showing survivor faces as goal increases
  • 50% increase in high-value fundraising goals
  • Addressed all 6 factors: emotional appeal, no typing, clear impact, one decision, temporal compression (immediate vs delayed impact)

Sketchnote by Chris Noessel

Designing for AI: UX Patterns, Practice, and Product Differentiation By Dave Brown

AI Design Pattern Framework Overview

• 5-phase framework for designing AI interactions: Discover > Instruct > Observe > Refine > Return
• Key challenge: moving from deterministic workflows to adaptive, real-time generated experiences
• Designer’s job: help users provide right context, surface reasoning, give control for iteration

Discovering Pattern

• Blank slate problem: users don’t know what AI system can do
• Common solutions: open input, suggested prompt cards, disclaimers, sparkle icons
• More advanced ways:
• Mode switching (plan/debug/ask modes, cycle effort controls)

Instructing Pattern

• Think about helping the users provide the right context into the system
• Examples:
• Multimodal input beyond natural language (LumaAI)
• Visual controls (Figma’s 2x2 matrix for copy tone)
• Spatial input (Visual Electric drawing placement)
• Context filtering (Writer.com’s graph-based filtering)
• Core UX principle: context is king - natural language alone often insufficient

Observing Pattern

• Show AI reasoning and processing steps vs. simple loading states
• Examples:
• Elicit: stepper pattern showing gather papers > screen papers > extract data > generate report
• Pencil.dev: agent canvas with sub-agents working visually
• Kimi: personified agent swarm with named roles
• Orbital Copilot: clickable citations linking to highlighted source PDFs
• Multimodal outputs: Runway generating color swatches + imagery + motion from single prompt
• Core UX Principle: surface the why - show reasoning, sources, processing time; help ser expectations and build trust in the AI output.

Refining Pattern

• Avoid long chat scrolling - enable quick inline adjustments
• Examples:
• Grammarly pattern: overlay suggestions with preview and insert
• Cursor: follow-up questions with 1/2/3 options
• Code diffs as refinement model for other domains
• Grok’s “think harder” - same prompt, different model, replaces output

Learning Pattern

• Memory management as user control issue
• Examples:
• ChatGPT model: default remember with manual deletion and incognito mode
• Personalization based on interaction patterns
• Pickle: knowledge graph visualization showing connections over time
• Gap: no multiplayer patterns yet, all single-user focused
• UX Principle: design for intervention - make it easy for users to find the outputs, give user control over the variations they want.

“So I think key job of the AI designer is helping users provide that right context, helping them surface to the user why something happened, and then giving them the level of control so they can iterate and refine that over time.”

Sketchnote by Chris Noessel

Beyond Experience: Who Benefits From the Systems We Design? By Ayesha Moarif

Speaker’s Journey & Core Thesis

• 10 years in UK public service design
• “Prototalk” - reflective piece on current thoughts about user-centered design limits
• Still believes HCD has value, but learned it’s not the complete answer for public services
• Key insight: HCD based on consumer model doesn’t fully apply to government services

Fundamental Differences: Public vs Private Services

• Public aren’t consumers - can’t opt out of taxation, immigration, education, healthcare
• Relationship shaped by obligation, power, eligibility - not just transactional
• Systems categorize users by risk/legitimacy, not just user satisfaction
• Example: Visa website had high user satisfaction but people chose wrong visas (rejection consequences)
• Less consumption power = more dependent on public services, more systems to navigate
• Friction often exists for legitimate reasons (legal obligations, protection, compliance)

Systemic Design Challenges

• We rarely design for actual system users. Frontline workers (nurses, caseworkers) rarely designed for - they’re actual system intelligence
• Staff jobs designed for compliance and replaceability, not human judgment
• “Are you conscious?” phone triage example - system designed for compliance, not logic
• Political changes can instantly make HCD work irrelevant (Trump administration example)
• UX can serve any agenda - not inherently good without explicit stance on power/purpose

Alternative Framework: Public Experience Design

• Focus on “How do we experience democracy?” rather than just user experience
• Design for agency: What role do citizens play? How can they use their data?
• Collective representation in systems (not just individual users)
• Transparency about service performance across neighborhoods/regions
• Friction as data - learning from system pain points
• Hilary Cottam’s approach: Design relationships and networks, not just products/services
• Better question: “How do we participate in public life given available technology?”

Sketchnote by Chris Noessel

The Wobbly Chair By Brandon Harwood

The Legibility Problem in Design

• German forest management analogy: 18th century foresters created systems to measure/optimize lumber yield
• Colored nails categorized trees by size and estimated value
• “Normal forests” - abstracted representations focused only on extraction value
• Diverse ecosystems replaced with monoculture rows of single species (poplars)
• Initial success followed by “waldsterben” (forest death) after ~80 years
• Vulnerability to wind, soil exhaustion, pest invasibility
• Required constant external support to mimic original ecosystem functions
• Parallel in UX: Making design “legible” to business through frameworks, KPIs, processes
• Double diamond, personas, OKRs created to justify design value
• Successful at gaining “seat at the table” but narrowed design scope
• Generalists pushed out for specialists (design systems, content, AI designers)

Crisis of Imagination in Current Design Practice

• Two concerning AI studies: Georgetown (homogenizes idea spaces), Oxford/MIT/CMU/UCLA (reduces individual capability after 10 minutes)
• Business leaders “do not fuck with complexity” - only deal with legibility
• Design becoming recipe-following vs. cooking from scratch
• Students increasingly frustrated without specific formulas for design problems
• Research findings ignored when they don’t align with leadership goals
• Homogenization visible everywhere (AI-generated content all looks the same)

Pragmatic Imagination as Solution

• Ann Pendleton-Jullian and John Seely Brown framework: Two modes of imagination
• Sense-making: Perception + reasoning (narrow gap, creates systems/heuristics)
• Sense-breaking: Speculation + experimentation + play (widens gap, disrupts norms)
• Successful examples of sense-breaking in business:
• Nick Foster’s “future mundane” approach → Head of Design at Google X
• IKEA insect meatballs speculation (2015) → plant-based meatballs with 4% carbon footprint
• IKEA catalog of near future → influenced actual catalog
• Actionable practices: Experimentation (prototyping for questions not answers), Play (unstructured exploration), Design fiction (tangible future representations)
• Strategy: Learn stakeholder language, make imagination results legible, hide methods if necessary
• Responsibility to maintain “underbrush” practices regardless of explicit permission

Sketchnote by Chris Noessel

Designing What Matters: Beauty, Data and the Human Touch in the Age of AI By Marianne Ashton-Booth

AI Era Design Challenges

• Design velocity paradigm increasingly prevalent - can test everything, generate anything, optimize relentlessly
• AI tools can produce thousands of interface variations in 3 minutes
• Risk: Building products faster through short-term learning choices without asking “what are we actually trying to create and why?”
• Delay between action and feedback creates problems (shower analogy - CRM notification strategy example)
• Senior leadership expectations changing as AI makes ideation/prototyping much faster

Memorable vs Forgettable Products

• Two valid approaches: consistency/predictability vs distinction/emotional connection
• Memorable products create connection, not just function
• Spotify reflects identity back to users
• Monzo uses tone/clarity to make money less intimidating
• Don Norman’s 3 levels: visceral (gut reaction), behavioral (does it work), reflective (what does this say about me)
• Aesthetic usability effect: people perceive attractive things as easier to use
• Research insight: Alzheimer’s patients lose cognitive function but aesthetic preferences remain - beauty sits deeper than cognition
• Peak-end rule: people remember peaks and endings, not every moment equally
• Royal Mail example: customer perception shaped more by friendly postman than operational efficiency

Practical Framework for AI + Design

• ITVX running 1-year “long holdout” experiment - segment gets no feature updates to measure long-term compound value
• Data tells what happened, not what mattered most or what felt right
• When everyone has the same AI tools, value shifts from ability to make to ability to recognize what’s useful
• Three key approaches:
1. Curate, don’t generate - AI creates options, humans choose what matters
2. Design, don’t default - some moments need intention beyond following patterns
3. Decide, don’t delegate - AI explores, humans make final calls

“ So remember, in a world where machines can produce endless outputs, the real differentiator isn't going to see speed or optimization, but it will remain as that. I believe it will make us the human ability to recognise what truly resonates. It's about your perspective, your understanding of what matters, even when that can't be measured. And that perspective, something that is becoming more and more valuable, not less.”

AI Stupefaction (And What We Can Do About It) By Christopher Noessel

AI Stupefaction Research

• Definition: AI stupefaction = systems that reduce cognitive ability by taking over thinking tasks
• From Latin: stupeo (stunned to silence) + faccio (to make)
• Different from innate ability - refers to circumstances that impair cognition
• Navigation study findings: 48% performance collapse when GPS removed after 4 days of use > AI causes stupefaction 
• Participants couldn’t navigate 12-decision route without assistance
• Measured decision time as proxy for confidence
• Effect size (Cohen’s D) of 1.09 = “overwhelmingly massive”
• Business reframe: Overreliance problem when AI is active
• Humans 6.4x worse when AI provides incorrect information
• More relevant to business than deskilling during downtime

Human-First Design Solution

• Modified interface results: 19% performance improvement (67% total swing from original)
• Removed blue line guidance, added post-decision feedback only
• Three feedback states: on track, alternate route, wrong way
• Participants learned route so well that checking reference slowed them down
• IBM Maximo application: Asset inspection workflow
• Human inspector identifies defects first, then compares with AI findings
• Three outcomes: both found it, human caught what AI missed, AI caught what human missed
• Micro-learning opportunities when AI catches human errors

Cognitive Forcing Functions

• Human Goes First pattern: Only method that requires skill practice while fighting overreliance
• User completes task while AI works in background
• Compare results for learning opportunities
• Maintains human expertise while leveraging AI assistance
• Key insight: Poorly designed AI causes stupefaction - well-designed AI can make humans smarter
• Designer leverage: Most positioned profession to implement these mitigations

. . .

👀 Stay tuned!

We’ll start releasing the videos from the Talks Day on our UXLx Videos page, gradually, when we launch the next UXLx event. In the meantime you can check hours of content from the previous editions.

Sign-up to our mailing list to get the latest UXLx news and updates.