Michael Bervell, who founded TestParty and has been a guest on Talk Commerce a number of times, published his company’s first-half numbers, and one of them is worth sitting with. After TestParty remediated accessibility issues directly in its customers’ source code, those customers saw a 14.25% average increase in AI-referred traffic, measured by comparing the three months before each launch to the three months after.
That figure is self-reported by a company that sells accessibility remediation, so treat it as a signal to test rather than settled law, but the mechanism behind it holds up, and the mechanism is the part worth your attention.
Screen readers and AI crawlers want the same thing
Accessible markup is structured markup. A screen reader needs a proper heading hierarchy so someone can skim a page, it needs alt text to describe an image, and it needs labeled fields and semantic elements so a person can tell a button from a link without seeing the screen. An AI crawler reading that same page needs those same things to work out what the page is about and to pull a clean answer out of it, so when you ship a vague pile of divs with no headings and images that say nothing, you make the page harder for a blind user and harder for ChatGPT in the same move. The alt text you write for someone who cannot see an image is the text a model reads to learn what the image shows, and the heading a keyboard user skims to is the heading an AI quotes when it answers a question.
Accessibility stopped being only a legal question
Accessibility lived in the legal column, where you did enough to lower the risk of a demand letter, often by dropping in an overlay widget that promised compliance without touching the underlying code. Those overlays sat on top of the page instead of fixing it, and they did nothing for how a machine reads your content. Fixing the source code is more work, and it is the version that pays you back in more than risk reduction, because the same edits that make a page usable make it legible.
The practical version
You can treat this as charity if you like, and the 1.2 billion people with disabilities are reason enough, but you can also treat it as discoverability. The work of making your site readable to a person using a keyboard or a screen reader is most of the work of making it readable to the agents doing product research on your customers’ behalf, so write proper headings, describe your images, label your fields, use semantic elements, and keep your words in the HTML instead of hiding them behind scripts. You end up with a site that is easier to use and easier to quote.
Where this lands for us
At Content Cucumber we say we are a company that employs humans to write content for humans, and accessibility is that sentence taken at its word, because “humans” includes the ones using assistive technology, and writing for all of them is the same craft that makes your content easy for AI to find and cite.
Let us help you get found in AI search results
AI Search Optimization →
This is a great point about accessibility becoming increasingly important for AI search. Clear content structure, descriptive headings, meaningful alt text, and accessible page elements can make content easier for both users and AI systems to understand. I also think businesses should treat accessibility as part of their broader technical SEO and content strategy rather than as a separate task. The connection between better user experience, clearer content signals, and AI visibility is definitely worth paying attention to.