Automated accessibility scans identify a specific category of issues: those that can be determined by evaluating HTML, CSS, and ARIA attributes programmatically. This covers approximately 25% of all Web Content Accessibility Guidelines (WCAG) conformance issues. The remaining 75% requires human evaluation to assess.
| Issue Category | What Scans Detect |
|---|---|
| Missing Text Alternatives | Images without alt attributes or with empty alt attributes where alternative text is required |
| Form Labeling | Input fields that lack associated label elements or accessible names |
| Document Structure | Skipped heading levels, missing document language declarations, and incorrect landmark usage |
| ARIA Attribute Errors | Invalid ARIA roles, missing required ARIA properties, and conflicting attribute values |
Code-Level Issues Scans Cover Well
Scans perform best when the determination is binary. An image either has an alt attribute or it does not. A form field either has a programmatic label or it does not.
A heading sequence either follows the correct order or it skips a level. This binary detection applies to a range of structural and semantic code patterns.
Missing document language, duplicate IDs, empty links, and buttons without accessible names all fall into this category. The scan reads the code, compares it against a defined rule, and flags anything that deviates.
Where Scans Reach Their Detection Boundary
A scan can confirm that an alt attribute exists on an image, but it cannot evaluate whether the text inside that attribute accurately describes the image content. This distinction applies across many WCAG criteria.
Scans cannot assess whether a custom interactive component works correctly with a keyboard. They cannot determine if content presented visually conveys the same meaning when read aloud by a screen reader. They cannot evaluate whether error messages provide sufficient guidance to correct a form submission.
These types of assessments require a human evaluator who can perceive the content, interact with it using assistive technology, and judge whether the experience meets the WCAG success criteria in context.
How Automated Scan Accessibility Issues Map to WCAG
The issues scans identify map primarily to WCAG criteria that have clear programmatic indicators. Criteria related to info and relationships (1.3.1), name, role, value (4.1.2), and parsing of HTML structure fall heavily into what scans flag reliably.
Criteria that require interpretation, such as whether content is presented in a meaningful sequence or whether instructions do not rely solely on sensory characteristics, fall outside scan detection. The overlap between what scans flag and what WCAG requires is real but limited to that 25% boundary.
Scans as a Starting Point
Automated scans are valuable for catching code-level automated accessibility issues at scale and on a recurring basis. They surface the issues that no human needs to spend time identifying manually. The 75% of WCAG criteria outside scan detection still requires a thorough audit conducted by a qualified evaluator.
Knowing what scans cover, and where they stop, sets accurate expectations for what any scan report represents.