Your accessibility widget did not stop the lawsuit
In 2025, 1,416 US businesses that had an accessibility widget installed were sued anyway. If you bought one and a demand letter still landed, you were not careless. You were sold a product that does not do what the page said it did.
This page is about what the widget actually fixed, what it never touched, and what a court does with the fact that you had one. It is written for the person holding a demand letter with a 30 day clock on it, and for the person who wants to not be that person.
The uncomfortable part first. Having an overlay installed is not neutral at trial. Plaintiffs use it to show you knew the obligation existed and chose a workaround instead of a fix. The subscription receipt becomes evidence of knowledge.
The numbers, and where they come from
| Finding | Figure | Source |
|---|---|---|
| US businesses sued in 2025 that had a widget installed | 1,416 | UsableNet tracker |
| Share of 2025 defendants already running some accessibility solution | 38.5% | AudioEye 2026 litigation report |
| FTC penalty against accessiBe, final April 2025 | $1,000,000 | FTC |
| accessiBe class settlement, paying out now | $1,200,000 | Court-approved settlement, 2026 |
| Typical overlay subscription | about $1,200/yr | Vendor published pricing |
Two of those deserve care. The 38.5% figure is a vendor's own litigation research rather than a court statistic, and AudioEye competes with the companies it is counting, so treat it as directional. The FTC order is not directional. It is a federal enforcement action.
What the FTC actually found
The order, finalised in April 2025, turned on advertising. The Commission's complaint was that accessiBe represented its automated tool could make a website WCAG compliant, and that it did not have reliable evidence to support that. The order carries a one million dollar payment and prohibits claims that an automated tool makes a site compliant without evidence behind them.
Read that back slowly, because it matters for your file. A federal regulator examined the claim "this widget makes you compliant" and concluded the evidence was not there. If you relied on that claim, you relied on something a regulator has since acted against.
The case that shows the shape of it
A small online florist, Bloomsybox, subscribed to UserWay's overlay in 2023 and was served with an ADA website accessibility lawsuit roughly six months later. The class action that followed cleared a procedural hurdle in February 2026. Its allegation is specific and worth knowing: the overlay had not fixed core problems such as missing alternative text and broken keyboard navigation, and the advertised legal protection did not prevent the suit.
Missing alt text and broken keyboard navigation are not exotic failures. They are the two most common findings in any manual audit, and they are precisely what an overlay is least able to repair, because both require knowing what an image means and how your interface is meant to be operated.
What an overlay can and cannot do
An overlay is JavaScript that loads after your page and tries to patch it in the browser. That constrains it absolutely:
- It can offer a toolbar for contrast, text size and spacing, and guess at some machine-detectable issues.
- It cannot know what a photograph depicts, so it cannot write meaningful alternative text. A generated description of "image" satisfies a scanner and tells a blind customer nothing.
- It cannot reliably fix keyboard traps or focus order, because those come from how your components are built.
- It cannot fix a checkout that fails on a screen reader, which is the flow that actually gets litigated, because that is where the money is.
- It cannot make a decision requiring judgement, such as whether an error message explains how to correct the error.
Screen reader users widely report turning overlays off. That matters commercially as well as legally: the tool meant to serve your customer is the thing they are switching off to get through your checkout.
If a demand letter has already landed
These typically give you 30 to 60 days. The order below is deliberate.
- Do not ignore it and do not reply off the cuff. Silence escalates it and an improvised response can concede things you did not mean to.
- Preserve the evidence of your current state. A dated audit of the site as it stands today is worth more than the same audit run after you start changing things.
- Baseline against WCAG 2.2 AA, automated scan plus a manual keyboard and screen reader pass. Automated tooling alone catches a minority of real barriers.
- Prioritise the flows that carry revenue rather than chasing a violation count: product page, cart, checkout, search, account.
- Fix in the code, as reviewable changes, so the fix survives your next deploy.
- Then publish an accessibility statement that describes what you did, when, and what is still outstanding. A dated, honest statement is evidence of good faith. An aspirational one is not.
Why this keeps happening after the fix
Accessibility is not a state you reach, it is a state you drift out of. Someone uploads a product image with no alt text on a Tuesday. A new component ships with a focus trap. A marketing page goes live with contrast that fails. Nothing in your process flags any of it, because nobody is employed to look every day.
That is the actual gap the overlay was sold into. The need is real. The product was wrong. What closes it is a check that runs every day against the live site and tells you the moment something that used to pass starts failing.
Where to start, free
- Run the accessibility check for an honest baseline of what is failing right now. Just your email address.
- Check which obligations apply to you, since ADA is rarely the only one in play. AODA, the Accessible Canada Act and the European Accessibility Act all have their own triggers.
- See what is due next if you are in Canada.
Common questions
Does having an accessibility widget protect me from being sued?
No, and the evidence runs the other way. UsableNet's tracker recorded 1,416 US businesses sued in 2025 that had a widget installed. Some vendors advertise legal support or a warranty alongside the subscription, but the Bloomsybox class action against UserWay alleges exactly that promise did not prevent an ADA suit landing about six months after the business subscribed. Treat a vendor guarantee as a commercial term between you and the vendor, not as a defence available to you in court.
Is an overlay worse than doing nothing?
For litigation risk, there is a real argument that it is, and it is not a comfortable one. Doing nothing leaves you with a site that fails. Installing an overlay leaves you with a site that still fails plus a documented, dated record showing you understood the obligation and bought a workaround. Plaintiffs use that record. It does not help you to have shopped for a solution if the solution did not work.
What did the FTC actually decide about accessiBe?
The order was finalised in April 2025. The Commission's position was that accessiBe represented its automated tool could make a website WCAG compliant without reliable evidence supporting that claim. The order requires a one million dollar payment and bars claims that an automated product makes a site compliant unless there is evidence behind them. It is an advertising enforcement action, so it does not itself rule on whether any particular website is accessible, but it is a federal regulator formally rejecting the central marketing claim of the category.
What does real remediation cost compared with a widget?
An overlay runs roughly 1,200 dollars a year. A manual accessibility audit for a small business is commonly quoted between 2,500 and 15,000 dollars, and AODA-specific audits for Ontario small and medium businesses more typically between 2,400 and 6,000. Remediation is usually quoted in a similar band to the audit. The honest comparison is not widget versus audit though, because they do not do the same thing. It is a subscription that does not fix the failures against one-off work that does, plus ongoing monitoring to catch the new ones.
Will an automated scan find everything?
No, and anyone promising otherwise is repeating the claim the FTC acted on. Automated tooling reliably finds machine-detectable issues such as a missing alt attribute, a form control with no label, or a contrast ratio below threshold. It cannot tell you whether the alt text is meaningful, whether the focus order makes sense, or whether your checkout can actually be completed by somebody using a screen reader. Automated scanning narrows the manual work. It does not replace it.