Scan-based platforms misguide compliance efforts because automated scans detect only about 25% of WCAG issues. The other 75% require human evaluation against the Web Content Accessibility Guidelines. When a platform reports a high accessibility score based on scan data, decision-makers assume the site is close to conformance. It is not. The score reflects what the scanner could check, not what the standard actually requires. This disconnect is how organizations end up with a dashboard full of green checkmarks and a website that still has significant accessibility issues.
| Factor | Scan-Based Platform | Audit-Based Approach |
|---|---|---|
| WCAG Coverage | Approximately 25% of issues | Full coverage of applicable criteria |
| Conformance Claim | Not possible | Supported by auditor review |
| Risk Signal | Misleadingly positive | Accurate picture of status |
| Evidence Value | Weak for legal defense | Strong documentation trail |
What scans actually detect
Automated scans are pattern matchers. They parse HTML, run rule sets, and flag what their rules can recognize. Missing alt attributes, empty form labels, color contrast ratios on text, and some ARIA issues fall within that window.
What sits outside that window is most of WCAG. Whether alt text is accurate. Whether a heading structure reflects the page’s real hierarchy. Whether keyboard focus order matches visual order. Whether a custom component announces its state to a screen reader. These require human evaluation.
A scan cannot read a page the way a user does. It reads the way a parser does.
How does a scan-based platform create false confidence?
The platform presents a score. The score reduces a complex standard to a number. Leadership sees 94 out of 100 and assumes the work is nearly done.
But the 94 is scored against the portion of WCAG the scanner can check. The 75% the scanner cannot check is invisible in the score. A site scoring 94 on a scan can still have dozens of serious issues a blind user would hit on the first page.
This is how scan-based platforms misguide compliance efforts at the executive level. The number looks like progress. The legal and user-impact reality tells a different story.
The documentation problem
Scan reports do not support a conformance claim. An accessibility conformance report is a review of every applicable WCAG success criterion, with a conformance level assigned to each. That review is done by an auditor, not an algorithm.
If a company tries to complete a VPAT using scan output alone, the resulting ACR is not credible. Procurement teams reviewing it will either reject it or flag it for follow-up. Legal teams reviewing it after a demand letter will find it thin.
The VPAT and ACR process depends on audit evidence. Scan data is not a substitute.
Risk exposure in litigation
Website accessibility lawsuits do not reference scan scores. Complaints identify specific issues on specific pages, usually surfaced by a plaintiff using a screen reader. Most of those issues are in the 75% category that scans miss.
A company relying on a scan-based platform may receive a demand letter describing issues its dashboard never flagged. Defense then has to explain why the platform showed a high score while the site was inaccessible to the claimant. That is a difficult position.
An accessibility audit produces the kind of documentation that actually supports a defense: a dated report identifying issues against WCAG criteria, followed by a remediation record.
Where scans do fit
Scans are useful for monitoring. After an audit identifies issues and remediation closes them, a scan can watch for regressions on the portion of WCAG it can detect. New content gets pushed, a developer misses an alt attribute, the scan flags it.
That is the correct role for scan technology. Ongoing checks on known-detectable issues, layered on top of audit-based conformance work. Not a substitute for the audit.
VPAT details are discussed further here.
How real AI fits into accessibility work
AI that claims to automate WCAG conformance is making a promise the technology cannot keep. The same limits that apply to rule-based scanners apply to AI scanners. Context, intent, and user experience are not reliably machine-readable at the depth WCAG requires.
Accessible.org Labs is actively researching how AI can make auditing and remediation more efficient for skilled practitioners. The framing matters. Real AI supports the auditor and the developer. It does not replace the evaluation.
FAQs
Can a scan-based platform get my site to WCAG conformance?
No. A scan detects approximately 25% of WCAG issues. Reaching conformance requires a human auditor reviewing every applicable success criterion, followed by remediation and validation. A scan-based platform can support monitoring after that work is done, but it cannot produce conformance on its own.
What should I look for in an accessibility platform instead?
Look for a platform built around audit data, with a process for tracking issues through remediation and validation. The Accessibility Tracker Platform organizes audit findings, supports remediation workflows, and generates progress documentation. The core record is the audit, not a scan score.
Why do scan-based platforms exist if they cannot determine conformance?
Scans are inexpensive to run at scale, which makes them attractive to package as a product. The issue is not the technology. It is how the technology gets positioned. When scan output is sold as a compliance status, buyers form inaccurate expectations. When it is positioned as regression monitoring, it serves a clear purpose.
Does a scan have any value before an audit?
It can give a rough sense of surface-level issues on a site, which is useful for early budgeting conversations. It should not inform the scope of remediation or serve as a basis for any conformance claim. The audit is where scope and status get determined.
The difference between a dashboard score and a conformance claim is the difference between a signal and a record. Compliance rests on the record.
Contact Accessible.org to discuss an audit-based approach for your site: Contact Accessible.org.
