Guides
A practical field guide to evidence-first usability assurance
Start with authority and a clear goal, then preserve the distinction between what the browser observed and what a synthetic AI User inferred.
AI usability testing: what it can prove and where people still matter
Understand AI usability testing, separate observations from interpretation, and choose the right evidence before releasing an AI-built app.
Open guideA SaaS onboarding UX checklist that goes beyond signup
Test the full SaaS onboarding journey, from a clear promise and account recovery to first value, saved work and returning users.
Open guideHow to write a Golden Journey Test Contract
Use this practical worksheet to define one critical user journey, its authorized actions, safe data, success criteria and release evidence.
Open guidePrivacy-conscious journey testing: collect evidence with less exposure
Plan safe test data, bounded capture, redaction, retention and sharing so a useful journey report does not become an unnecessary data archive.
Open guideHow to check whether a UX improvement actually worked
Turn a UX finding into a testable improvement, compare equivalent journeys, and distinguish observed progress from unproven conversion claims.
Open guideYour AI-built app works in the demo. What should you test before release?
Review an AI-built app beyond its demo: fresh accounts, saved work, recovery paths and the evidence needed for a bounded release decision.
Open guideSmoke testing, regression testing and usability testing answer different questions
Learn how smoke, regression and usability testing contribute different evidence, and combine them around one release-critical user journey.
Open guideBuild a user-journey coverage matrix without testing everything
Map critical tasks, roles, starting states and recovery paths into a small coverage matrix that makes tested and untested journeys visible.
Open guideWrite usability test tasks that do not reveal the answer
Write goal-based usability tasks that reveal navigation and comprehension problems without teaching the tester which buttons to click.
Open guideWrite a UX bug report another person can reproduce
Turn vague UX feedback into a reproducible report with starting conditions, observed behavior, task impact, evidence and a clear repair check.
Open guidePrioritize UX friction without hiding behind a score
Prioritize usability findings by task impact, evidence quality, workarounds and release context instead of treating a single score as the decision.
Open guideEmpty-state UX: test the first useful action
Test empty screens for a clear explanation, an achievable next action and truthful state handling before the first customer record exists.
Open guideForm validation UX: help people recover without starting over
Review form errors as a recovery journey: identify the problem, explain the correction, preserve appropriate work and verify the final result.
Open guidePassword reset is a product journey, not just an email
Review password recovery from the first request through email, link handling, confirmation and a successful return to the intended product task.
Open guideTest invitation acceptance for new and existing users
Review invitation acceptance across new accounts, existing sessions, expired links and workspace selection without sending test invites to real customers.
Open guideWorkspace switching: make the active organization unmistakable
Test organization switching, saved links and role changes so users can identify their active workspace before viewing or changing business data.
Open guideTest multi-role handoffs without losing the user's place
Review requester-to-reviewer handoffs for clear ownership, current versions, actionable return paths and a visible record of what happens next.
Open guideMobile web testing should follow a task, not a screenshot
Test mobile web journeys through navigation, input, error recovery and saved results instead of checking only whether a page fits a narrow screen.
Open guideA keyboard walkthrough for your most important user journey
Review a critical task using keyboard input, checking visible focus, usable controls and recovery while keeping accessibility claims within the evidence.
Open guideLoading, retry and uncertain saves: test the waiting experience
Review asynchronous tasks for clear progress, honest failure states and safe recovery when users cannot tell whether an operation completed.
Open guideSearch and filter UX: test the path back from no results
Test search and filtering with known records, active constraints and no-result recovery so users can understand and change the view they are seeing.
Open guideFile upload UX: the journey continues after the progress bar
Test file selection, limits, processing, failure recovery and saved attachments so an upload's visual success matches a useful final result.
Open guideA report export is complete only when the file is useful
Test report exports for the intended scope, understandable progress, usable file contents and a reliable path from on-screen filters to shared output.
Open guideTest pricing, trials and plan limits as one continuous journey
Review whether pricing promises, trial setup and in-product limits tell a consistent story before asking users to upgrade or commit to a plan.
Open guideTest an AI feature's promises, uncertainty and user control
Review an AI feature's expectations, evidence, correction path and action boundaries without treating model output as proof of accuracy or user trust.
Open guideReview and approval UX for AI-generated content
Test the path from AI-generated draft to human review, revision and approved action while keeping versions, destinations and responsibilities clear.
Open guideWhat an agency should include in a UX handover pack
Prepare a client handover that connects tested journeys, known limitations, repair checks and release ownership without claiming more than the evidence shows.
Open guideTurn subjective UX feedback into acceptance criteria
Translate requests such as clearer, easier and more polished into observable task outcomes while preserving questions that still require human research.
Open guideRun a weekly journey review that produces decisions
Keep a weekly UX review focused on changed journeys, current evidence, unresolved findings and named actions instead of accumulating screenshots and scores.
Open guideTurn customer-support reports into repeatable journey tests
Convert support symptoms into safe, reproducible journey checks while preserving uncertainty, protecting customer information and verifying the actual repair.
Open guideWrite a bounded Test Contract
Define allowed targets, identities, actions, budgets, side effects and stopping rules before execution.
Open guideRead a Circuit report
Separate browser observation from AI interpretation and verify the evidence behind severity.
Open guidePrepare domain verification
Issue and publish the exact DNS TXT proof required before a target can become executable.
Open guideChoose a retention mode
Balance collaboration and regression needs with short-lived raw evidence and deletion controls.
Open guideBuild a regression gate
Approve a Baseline and compare new, persistent and resolved Findings after a product change.
Open guideEvaluate an AI-built app
Cover the release-critical journey, generated copy and non-happy-path interface states.
Open guide