What an automated scan can tell you
Automated testing is useful for finding repeatable problems across many pages. A tool may flag missing accessible names, certain contrast failures, and invalid markup. Run scans during development and use them to catch regressions. Review each result before assigning work: a flagged item can require context, and a clean scan does not mean that a page is accessible.
What requires a person to evaluate
People need to assess whether instructions are understandable, alternative text conveys the purpose of an image, focus moves logically, and a task can be completed with assistive technology. These questions depend on the content and interaction. A button can have an accessible name yet still fail to explain its action. Our WCAG accessibility audits combine tool output with these manual checks.
Test workflows, not just URLs
For a checkout, test product selection, quantity changes, address entry, validation, payment, and confirmation. For a portal, include sign-in, navigation, downloads, and account recovery. Record the state of the interface when an issue occurs. A screen reader announcement that is missing after a submission will not necessarily appear in a scan of the initial page. Use the audit preparation guide to define representative tasks and access.
Make keyboard and screen reader checks distinct
Keyboard testing asks whether someone can reach and operate controls, see where focus is, and leave an interaction without becoming trapped. Screen reader testing also examines names, roles, states, structure, and announcements. Passing one set of checks does not prove that the other passes. Select browser and assistive technology combinations that fit the product, then record them in the test scope.
Turn results into a usable issue list
Deduplicate repeated component defects while preserving examples of where they occur. Record reproduction steps, expected and observed behavior, affected users, and the relevant success criterion. Explain whether the problem belongs to shared code, page content, or a third-party component. This makes remediation planning more useful than a long export of scan warnings.
Choose the right evaluation for the decision
A quick review helps a team locate obvious barriers. A scoped audit supports a more detailed remediation effort. A conformance report requires evidence against the applicable criteria and a clear account of coverage and limitations. If the buyer needs an ACR, discuss VPAT testing and reporting before testing starts. Do not present an automated score as a conformance certification.
Frequently asked questions
Why can a website pass a scan and still be difficult to use?
A scanner evaluates the rules and states it can inspect. It may confirm that a control has a name without determining whether that name explains its purpose, or miss a dialog that appears only after a particular action. Manual evaluation follows the task and examines context. A passing scan therefore does not establish that all relevant interactions or WCAG requirements have been checked.
Is there a reliable percentage of accessibility issues that automation finds?
There is no single percentage that describes every tool, product, and testing approach. Results depend on the rules implemented, the content and states reached, and whether the measurement counts criteria, defects, or repeated instances. Evaluate what the tool checks and what needs human review. A headline detection percentage is not a substitute for a documented coverage plan.
Does testing with a screen reader cover every WCAG requirement?
No. Screen reader evaluation is one part of a broader assessment. Keyboard operation, visual presentation, text enlargement, pointer interactions, timing, and other requirements may need separate checks. The relevant methods depend on the content and functionality. A manual WCAG audit should identify the coverage instead of treating one assistive technology as a complete test.
Where should automated checks fit into a release process?
Use them for repeatable checks during development and regression testing, with manual review for new components and important workflows. Investigate failures, document exceptions, and make sure the tests reach the relevant states. Keep responsibility for human checks explicit. The accessibility testing checklist provides a starting point for planning those reviews.