Accessibility overlays: the 2025 criticisms explained

Actualizado: 2026-05-03

On June 28, 2025, the European Accessibility Act entered full force for the private sector, and with it an avalanche of commercial offerings promising immediate compliance in exchange for adding a line of JavaScript to a website. Three months later, the landscape is revealing: lawsuits against companies using overlays have grown, disabled-user groups have issued increasingly pointed manifestos, and several European regulators have published guidance clarifying that these products do not constitute compliance by themselves.

Key takeaways

• An accessibility overlay is a JS script that dynamically modifies the DOM to supposedly improve accessibility — the best known in Europe are accessiBe, UserWay, EqualWeb, and AudioEye.
• The European Commission published in July 2025 an EAA clarifying note stating compliance must be substantive, not symbolic — solutions masking structural problems by modifying the DOM do not constitute compliance.
• Problems reported by real users are concrete: they overwrite JAWS/NVDA configurations, generate incorrect alt text in critical contexts, and the user panel itself is often not accessible.
• The Murphy v. Eyebobs case in the US established that accessiBe does not substitute for accessible design.
• The alternative is not magic but it is known: manual audit, substantive code correction, linter integration (axe-core) in CI, and an honest accessibility statement.

What overlays are and what they promise

An accessibility overlay is a JavaScript script added to a site that dynamically modifies content to supposedly improve accessibility. The best-known products in the European market are accessiBe, UserWay, EqualWeb, and AudioEye. Their promise is attractive: for a relatively low monthly fee, the tool analyzes the page, fixes issues automatically, adds a user options panel, and declares compliance with WCAG 2.1 or 2.2 at level AA.

The internal mechanics vary by product, but the common pattern is a detection engine that identifies potentially problematic elements and applies fixes like adding ARIA attributes, adjusting contrast, injecting auto-generated alt text, tweaking font sizes, or altering keyboard behavior.

The sales argument is clear: regulatory compliance without touching site code or redesigning components. For companies with legacy sites, limited budget, and a legal deadline approaching, the temptation is enormous. And yet, the evidence of recent years points to the promise not holding up.

Why the affected community rejects them

The strongest criticism comes from disabled users who actually rely on screen readers, magnification software, keyboard navigation, or other assistive technologies. In 2021 the Overlay Fact Sheet statement was published, signed by hundreds of professionals and activists, rejecting the model. The rejection has consolidated and widened since.

Reported problems are concrete:

• Overwriting user configurations: an NVDA or JAWS user with their profile tuned over years suddenly gets a layer that changes ARIA tags unpredictably, causing confusion.
• Incorrect alt text in critical contexts: auto-generated alt attributes using AI still produce wrong descriptions in banking forms or medical information.
• The panel itself is not accessible: many disabled users report that the overlay user panel has controls that do not receive focus, uses poor contrast, or imposes a navigation flow that breaks the user’s own. The frustration is double: they are sold a solution and then the solution adds barriers.
• Privacy: some overlays run detailed behavioral analysis to adapt profiles automatically, including data that can reveal the user’s disability. This raises GDPR concerns without explicit consent.

Legal ground has hardened

In the United States, ADA lawsuits against companies using overlays as their sole measure have grown. In several known cases, courts have ruled that using an overlay does not absolve the company of liability if the underlying site remains inaccessible. The most cited ruling is Murphy v. Eyebobs, where the court explicitly stated that accessiBe does not substitute for accessible design.

In Europe, the situation in September 2025 points in the same direction. The European Commission published in July 2025 a clarifying note on the EAA stressing that compliance must be substantive, not symbolic, and that solutions dynamically modifying the DOM to mask structural problems do not constitute compliance. Several national bodies, like Spain’s Centro de Referencia Estatal de Autonomía Personal, have published guidance explicitly discouraging overlays as a sole measure.

The practical consequence is that a company relying exclusively on an overlay to comply with the EAA is exposed to fines and civil suits with few defenses.

Why companies still buy them

Despite all of the above, overlays keep selling. The reasons are understandable if not justified:

• Information asymmetry: the executive team deciding to buy rarely has direct contact with disabled users or accessibility auditors, and the vendor’s marketing page looks convincing.
• Deadline pressure: with the EAA in force since June 2025, many companies discovered late they were not compliant and sought quick solutions.
• Apparent low cost: a monthly subscription looks cheap compared to an audit, code restructuring, and internal training.

The problem is that all three reasons dissolve under careful examination. Information asymmetry is solved by consulting associations like ONCE, CERMI, or equivalents in other countries, which publish clear positions. Deadline pressure is not eased by an overlay that legally does not comply: it only postpones the problem. And the real cost of an overlay, adding years of subscription to the audit that inevitably arrives, usually exceeds the cost of doing things right from the start.

What to do instead of an overlay

The alternative is not magic, but it is known and it works:

First step: a manual audit done by accessibility professionals, ideally with disabled-user participation in testing. This identifies real problems, prioritized by impact. It is not cheap, but it is the foundation of any serious plan.

Second step: substantive code correction. Proper contrasts in CSS, correct HTML semantic structure, ARIA attributes only where they actually add value, visible focus, coherent tab order, forms with associated labels, and systematic keyboard testing. This work happens on the site, not on top of it. It is the only work that produces real compliance.

Third step: integrating accessibility into the continuous development flow. Linters like axe-core integrated in the CI pipeline, accessibility reviews in pull requests, periodic team training, assistive-tech testing as part of QA. This prevents regression and turns accessibility from a point project into a sustained practice.

Fourth step (optional but valuable): publishing an honest accessibility statement. Admitting known issues not yet resolved, indicating a timeline, offering a contact channel for users encountering barriers. This transparency is legally useful under the EAA and builds trust with the community.

My read

After years watching this debate, my read is that accessibility overlays are a textbook case of an apparent solution that worsens the problem. They sell to those who do not know enough to distinguish, give a false sense of compliance, disincentivize real fixes, and actively harm the users they claim to protect. That they remain a profitable business says more about the information vacuum in the market than about their effectiveness.

My practical recommendation for anyone considering buying an overlay is very clear: do not. Invest that money in a real audit, internal training, and a correction plan on a reasonable horizon. It is more work, it costs more upfront, and it does not produce a solution in a week. But it is the only thing that meets the law, the only thing that improves user lives, and the only thing that holds up over time.

If you already bought an overlay before this information reached you, do not panic: use the current subscription as a temporary window while you prepare the transition to real work. But start the transition now, because when the first lawsuit or the first inspection arrives, no vendor marketing page will defend you.