In many teams, accessibility testing still happens just before release, usually during QA. Automated checks are run on selected screens, issues are recorded, and fixes are planned around the release timeline.

This setup catches basic gaps, but it does not reflect how the product behaves when someone actually uses it end-to-end.