Accessibility Overlays Don't Make Your Site Compliant

Each procurement cycle, the same pitch returns: drop a line of JavaScript on the site, become accessible, become compliant. Source-code accessibility looks unnecessary.

What overlays actually do, and why both parts fail

An overlay is a third-party JavaScript widget. It does two things at page load: presents an accessibility toolbar (contrast, font size, read-aloud etc.), and auto-repairs accessibility issues in the underlying code.

The toolbar

The widget duplicates features the visitor already has. Contrast, text size, magnification and read-aloud are built into iOS, Android, Windows and macOS, and into every major screen reader. A user who needs them sets them once and has them everywhere.

From the Overlay Fact Sheet:

"if those features were necessary to use the website, they'd be needed for all websites that the user interacts with. Instead, the widget is — at best — redundant functionality with what the user already has."

The widget overrides those OS preferences per site. A one-time setup becomes a per-site chore. It fails three WCAG criteria itself: 2.1.1 (keyboard), 1.4.13 (hover and focus content), 2.4.7 (focus visible). It adds barriers in the name of removing them.

UsableNet's 2026 Midyear Report names the outcome directly:

"For someone who relies on assistive technology, an inaccessible site with a widget is often harder to use than one without."

In WebAIM's 2021 Practitioner Survey, only 3.3% of respondents rated overlays very effective. An independent 2026 ACB study reported 2.4%.

The auto-repair

Automation has a ceiling. Per the W3C ACT Rules Framework, about 32% of WCAG 2.x A/AA Success Criteria can be tested fully automatically. The remaining ~68% need human judgement: alt-text quality, label clarity, focus order, error recovery, ARIA semantics.

Karl Groves:

"There's just too many things in Accessibility that are too subjective and too complex for a tool to test with enough accuracy to be considered a judgment."

What cannot be tested cannot be repaired. Any repair gets undone the moment a React, Angular or Vue component re-renders. In Groves' 2026 React test, an overlay reliably fixed only 46% of intentional accessibility issues across 29 components. The WebAIM Million 2026 found WCAG failures on 95.9% of the top one million homepages. Calling it AI does not raise the ceiling.

Nielsen Norman Group's Tanner Kohler walks through the same pattern from the user-research side:

The compliance promise is false

Liability sits with the operator across all three jurisdictions that matter for European buyers: Germany's BFSG (since June 2025), the US where the FTC fined a leading overlay vendor $1 million in January 2025 and UsableNet counted 831 ADA lawsuits against widget-equipped sites in H1 2026, and the EU where the EmpCo Directive takes effect 27 September 2026 with no transition period, banning unsubstantiated compliance claims.

Dennis Deacon names the inversion in a July 2026 essay:

"The icon installed to protect your organization from litigation has become a flag that invites it."

Across regulators and the accessibility profession, the verdict is the same: an overlay does not make an inaccessible site compliant. Germany's BfIT-Bund and 16 Länder authorities, who enforce EN 301 549 and BFSG, put it in regulatory language:

"Eine nicht-barrierefreie Website wird durch den Einsatz eines Overlay-Tools nicht zwingend barrierefrei nach den gesetzlichen Anforderungen."

The Overlay Fact Sheet, signed by 1,030+ accessibility professionals, puts it more bluntly:

"No overlay product on the market can cause a website to become fully compliant with any existing accessibility standard and therefore cannot eliminate legal risk."

The EU Directive 2024/825 "EmpCo" - which stands for "Empowering Consumers for the Green Transition" - gives the consensus enforcement teeth. The pitch that an overlay can "make any website compliant with WCAG" — the exact claim the FTC ruled deceptive in its January 2025 accessiBe order — is what its black list bans. The ban applies automatically, with no case-by-case review. A competitor or consumer organisation can file an Abmahnung. tollwerk lays out the German legal argument.

Why the message still sells

The consensus is near-total, yet procurement teams still buy. The vendor message reaches them through trusted disability-sector channels.

In an April 2026 open letter, Domingos de Oliveira documents a disability-community news outlet republishing overlay-vendor copy uncritically, with comments disabled. A major disability charity did the same.

In procurement, a disability organisation's endorsement often outweighs a fact sheet from consultants.

The pitch also mutates. A July 2026 newsletter by netz-barrierefrei.de tracks overlay firms pivoting into automated PDF remediation and "Abmahnschutz" (protection from accessibility lawsuits). The same compliance-in-a-click promise gets cloned onto adjacent products.

What to do instead

What replaces the overlay pitch is a way of working, not another product.

1. Design and code for accessibility from the start. Semantic markup, focus management, tested contrast. Decisions made at the code level, not injected at runtime.
2. Involve disabled users in research and testing. The people whose experience an overlay claims to improve are the ones who can tell you whether it did.
3. Test accessibility throughout the lifecycle. WCAG criteria in each sprint's Definition of Done. Regular audits with automated scans, manual review, and assistive-technology testing.
4. Build accessibility skills, or contract them in. IAAP CPACC and WAS give an in-house team a shared vocabulary. Specialist agencies and disability consultants fill gaps where staff cannot.
5. Publish an honest accessibility statement. Required under BFSG. Name conformance level, known gaps, and how to reach you.

A widget cannot change source code. The buyer who treats it as a shortcut keeps the liability, loses the budget, and excludes the users the product was sold to help.