A scan cannot replace an audit. Automated scans detect approximately 25% of accessibility issues by checking HTML, CSS, and ARIA attributes against a fixed set of rules. The remaining 75% requires human evaluation, which is what an audit provides.
Scans are pattern matchers. They flag missing alt attributes, empty form labels, and other code-level issues that can be detected programmatically. What scans cannot do is judge whether alt text is accurate, whether a form label describes its field correctly, or whether a screen reader announces content in a logical order.
What an Audit Covers That a Scan Does Not
An audit is a human evaluation of a product against WCAG success criteria. It includes screen reader testing with NVDA, JAWS, and VoiceOver, keyboard testing, visual inspection, code inspection, and browser zoom at 200% and 400%. An auditor identifies issues a scan has no way to see.
Treating a scan as an audit leaves three out of every four issues undetected, which is why WCAG conformance claims depend on human evaluation, not automated output.