Automated vs manual accessibility testing: what each one catches
Automated tools find many accessibility problems quickly, but not all of them. Where the line is, and how to combine both in a service clients can rely on.
Rongu team · Published
Automated accessibility testing is fast, repeatable and cheap to run on every page. It is also limited. Agencies that understand the limits can sell monitoring honestly and plan the manual work that remains.
What automation does well
Rule-based checks are reliable for problems that can be decided from the page's code and rendering:
- missing text alternatives on images and form controls,
- insufficient colour contrast for text,
- missing or duplicated IDs and broken ARIA references,
- empty links and buttons,
- missing page language and title,
- form fields without labels.
Run on a schedule, these checks catch regressions within days of a release, across every page.
What still needs a person
Some success criteria require judgement:
| Question | Why tools can't answer it |
|---|---|
| Is the alt text meaningful? | A tool sees that alt text exists, not whether it describes the image. |
| Can everything be done with a keyboard? | Focus order and traps depend on interaction. |
| Do headings describe their sections? | Structure is checkable; meaning isn't. |
| Are error messages helpful? | The wording and timing matter. |
| Does content reflow at 400% zoom? | Layout behaviour needs a look. |
Combining both
A practical approach for agencies:
- Automated scans on every client site, weekly or after each deploy.
- A manual checklist per site, reviewed at launch and after major redesigns.
- One report that shows both, so clients see the whole picture.
AccessWatch tracks manual checks next to the scan results and includes both in reports and accessibility statements.