A reproducible finding
- An action, expected result and observed result are recorded.
- A replay or browser measurement supports the specific claim.
- Screenshots identify the route, viewport and state involved.
Guide · UI and UX review
Find UI issues and broken user journeys
A working button does not tell you whether the next screen makes sense. Review the whole journey: the labels people follow, the choices they make, the feedback they receive and the way they recover from errors. Use the checklist below for a manual pass or to scope a Sybqa run.
Facts checked
Start with one task and an expected outcome
- Choose one task, such as finding a product and adding it to a cart. Write down the start page and the visible result you expect.
- Use synthetic data in a test environment. Include a normal path, an invalid input and a return to an earlier screen.
- Record each visited route, viewport, action and result. A route you discovered but never opened is still untested.
Compare screens and transitions
| Check | Example concern | Evidence to collect |
|---|---|---|
| Repeated controls | The same account control uses a different label or icon on another page. | Capture both controls, their accessible names and destinations. Confirm they have the same function before calling them inconsistent. |
| Labels and layout | A field label disappears after typing, or help text overlaps a control on a narrow screen. | Capture before and after at the same viewport; include the field and surrounding content. |
| Next steps | A completed step lands on a page with no clear way to continue. | Record the preceding action, destination and available controls. Treat unclear wording as a review lead until assessed in context. |
| Feedback | Submitting a form leaves the screen unchanged with no visible status. | Record the action, elapsed time, visible response and available network result. Silence alone does not prove the server failed. |
| Error recovery | Invalid input shows an error without a useful correction, or retry clears unrelated fields. | Record the synthetic input, exact error, correction and retry result. |
| Return visits | Going back loses a selection or contradicts an earlier confirmation. | Capture earlier and later states in the same journey. Replay to distinguish a reproducible defect from an uncertain observation. |
WCAG guidance addresses consistent identification of recurring components, suggestions for correcting detected input errors, and status messages that assistive technology can announce. These are useful checks within a wider UX review; this checklist does not establish WCAG conformance.
Separate a reproduced bug from a UX concern
A concern needing review
- A label seems confusing, but the intended business meaning is unknown.
- Screens differ, but their tasks or states may explain the difference.
- A screenshot suggests lost state, but the transition has not been replayed.
For example, a Save button returning a 404 is observable browser behavior. A label that might confuse a first-time visitor needs context and may need participant testing. Record both accurately instead of giving them the same verdict.
Use Sybqa to collect the evidence
Give Sybqa a link and a bounded goal, such as: “Check product selection through cart review at desktop and phone widths. Inspect labels, validation and feedback. Stop before payment.” Review the generated plan before running it.
Sybqa uses manual-testing techniques, AI-assisted visual review and real browser checks. The free plan covers browser checks without AI review. Paid AI review can surface UI inconsistencies and flow concerns across captured screens. Measurable claims are checked in the browser; subjective suggestions remain marked for review. See pricing for current allowances.
Sybqa is software, not a staffed manual testing service. It does not establish human-equivalent judgment, user satisfaction, backend persistence or complete coverage. A blocked or untested step never counts as passed.
Common questions
Can a page pass its checks while the journey still has issues?
Yes. Loading successfully or passing a field assertion covers a specific condition. Review transitions, labels, feedback and recovery across the whole task as well.
Does different wording always mean an inconsistency?
No. Confirm that recurring controls perform the same function in comparable contexts. Different tasks or states may justify different wording.
Can Sybqa replace testing with real users?
Browser checks and AI-assisted review collect evidence and possible concerns. They do not establish what real participants understand, feel or need. Use participant testing for those questions.
Sources
Related
Try Sybqa on your site
Explore bugs, UI consistency and user journeys from a reviewed plan. Inspect browser evidence and visual review leads in the report. The free plan gives 10 runs a month without AI review after you sign up. See pricing.