UXAtlas Journal

Field notes, methods, and sharp thinking for research teams.

A publication built from 93 research methods, practical playbooks, and essays on what actually works in the field.

The Accessibility Crisis in AI-Powered Interfaces: What WCAG Does Not Cover

Screen readers cannot parse streaming text. Adaptive layouts break spatial memory. Voice-first AI assumes everyone can hear. The accessibility frameworks we rely on were built for a world that no longer exists — and disabled users are paying the price.

March 2, 2026·23 min read

LATEST

New pieces

Browse archive

AI UX Engineering20 min read

Graceful Degradation: Designing AI Features That Fail Without Breaking Trust

Every AI feature will fail. Models hallucinate, services crash, confidence drops below usable thresholds. The question is never whether your AI will fail — it is whether users will still trust your product afterward. That depends entirely on how you design the failure.

AI UX Engineering21 min read

Latency Is a UX Problem: Engineering Perceived Performance When AI Models Think

The 100-millisecond rule that governed web performance for two decades breaks down when your backend needs three seconds to think. The teams winning the AI UX race are not making models faster — they are making waiting feel like progress.

Research Methods21 min read

AI-Assisted Qualitative Analysis: When to Trust the Machine With Your Research

AI can cut qualitative analysis time by 80%. It can also introduce systematic biases that poison your findings. The difference is not the tool — it is knowing precisely which analytical tasks to delegate and which to protect.

UX Research Methods25 min read

Beyond Nielsen’s 10: Usability Heuristics for the AI Era

Nielsen’s heuristics were built for buttons and menus. AI products need heuristics for trust calibration, graceful error recovery, and the strange new problem of systems that are confidently wrong. Here are eight principles to evaluate what the original ten cannot.

AI UX Engineering22 min read

The Architecture of Undo: Building Reversible AI Actions in Production Systems

Your AI agent just booked the wrong flight, sent a premature email, and modified a production database. Ctrl-Z does not work here. Reversible AI actions require event sourcing, compensating transactions, and an entirely new engineering discipline for undoing the real world.

METHOD DIGEST

Six methods worth revisiting

Full index

Participatory

Giga-Mapping

Giga-Mapping creates large-scale visual maps of complex systems, revealing hidden relationships and dependencies across multiple layers.

2–4 hours for initial session, ongoing refinement

Survey

Questionnaire

Questionnaires collect standardized, quantifiable data from large audiences through structured written questions about behaviors, preferences, and attitudes.

A week or more.

Observational

Mystery Shopping

Mystery Shopping sends trained evaluators to experience a service as real customers, revealing gaps between intended and actual customer experience.

1 day or more.

Participatory

Design Studio Method

Run a Design Studio workshop to rapidly generate, critique, and iterate on design ideas through structured sketching rounds with your team.

4 hours or more.

Observational

Bubble Test

The Bubble Test uses speech bubbles and images to uncover participants' emotional reactions, internal dialogue, and unfiltered opinions about designs.

30 minutes.

Participatory

Parallel Design

Parallel Design has multiple designers independently create solutions to the same problem, then combines the strongest ideas into one concept.

8 hours or more.

RESEARCH BEATS

Explore by goal

Browse all methods

Problem Discovery Testing & Validation Feedback & Improvement Generate Ideas Design & Prototyping Planning & Analysis Visualization & Communication