Accessibility isn't a checklist, it's a design review

Automated accessibility tools are useful and catch real problems: missing alt text, insufficient color contrast, unlabeled form fields. They're also frequently the only accessibility work a project does, run once before launch, treated as a compliance checkbox rather than a design input.

What automated tools structurally can't catch: whether a focus order makes logical sense to someone tabbing through with a keyboard, whether an interactive pattern is actually operable without a mouse, whether a "skip to content" link exists and is the first focusable element on the page. Those are design decisions, not lint errors.

The teams that ship genuinely accessible sites build it into design review the same way they'd review typography or spacing: does this component work with a screen reader, does this color pass contrast at the size it's actually used, does this modal trap focus correctly, asked during design, not discovered in a Lighthouse report after launch.

An automated score of 100 tells you the obvious problems are fixed. It doesn't tell you whether a real person using a keyboard or a screen reader could actually complete the task the page exists for.