Accessibility audits with AI: WCAG and the European Accessibility Act
An accessibility audit checks a website or app against standards such as WCAG to find barriers for disabled people. The European Accessibility Act has applied since 28 June 2025 to many consumer digital products and services sold in the EU, including e-commerce, banking and transport, with technical requirements aligned to WCAG through the EN 301 549 standard. Automated and AI tools can find many issues quickly, such as missing alt text, contrast and labels, but manual testing with keyboards, screen readers and disabled users is still needed.
What does the European Accessibility Act require?
The European Accessibility Act (EAA) is an EU directive that sets accessibility requirements for many products and services, and it has applied since 28 June 2025. It covers areas including e-commerce, consumer banking, e-books, transport ticketing and certain communication services, with some exemptions for micro-enterprises providing services. In practice, the technical benchmark for websites and apps is the harmonised standard EN 301 549, which incorporates WCAG success criteria. Enforcement and penalties are set by each member state. Businesses outside the EU that sell into it can also be affected. Details are on the European Commission’s EAA page.
What is WCAG?
The Web Content Accessibility Guidelines (WCAG), published by the W3C, define how to make digital content perceivable, operable, understandable and robust for disabled people. Version 2.2 is current, and level AA is the usual target. Our guide to accessibility in UI/UX design explains the principles in design terms.
What can automated and AI tools do?
AI adds value on top of classic scanners: grouping duplicate issues, explaining them in plain language, suggesting code fixes, drafting alt text for review and prioritising by impact on key journeys. It does not replace judgement. Automated checks are generally understood to find only a portion of real-world accessibility problems, so treat a clean scan as a starting point, not proof of compliance.
How do you run an AI-assisted audit?
- Define scope: List key journeys: sign-up, search, checkout, account, support.
- Run automated scans: Scan templates and key pages; export results.
- Use AI to triage: Group duplicates, explain issues and rank by severity and journey.
- Test manually: Keyboard-only, screen reader, zoom to 200% and reduced motion on each journey.
- Test with disabled users: Even a few sessions reveal problems no tool finds.
- Fix and retest: Fix in the design system first so fixes spread everywhere.
- Document: Record findings, fixes and an accessibility statement.
What should you fix first?
Why fix it in the design system?
Most accessibility issues repeat across a product because they live in shared components: a button without a visible focus state, an input without a label, a colour token with low contrast. Fixing the component fixes every instance. That is also how you keep AI-generated screens accessible, since tools reuse your components; see AI-ready design systems.
What does this look like in practice?
An online retailer selling into the EU runs an automated scan across its templates and finds 42 issue types, mostly low-contrast text, unlabelled inputs and missing alt text. An AI assistant groups them, explains each and suggests fixes. Manual testing then reveals what the scan missed: a keyboard trap in the checkout’s address lookup and a size selector that screen readers announce only as “button”. The team fixes the colour tokens and form components in the design system, rewrites alt text for key product images with human review, and repairs the checkout. Testing with two screen-reader users confirms the journey now works, and the company publishes an accessibility statement.
How do you stay accessible after the audit?
- Add automated accessibility checks to your build pipeline
- Include keyboard and screen-reader checks in design and code review
- Train designers, developers and content editors on the basics
- Re-audit key journeys after major releases
- Provide a way for users to report accessibility problems
Accessibility is a practice, not a one-off project. Teams that build it into their design system, reviews and release process find each audit smaller than the last, and products become easier for everyone to use, not only disabled people. That also reduces legal risk as enforcement grows.
Start with your most important journey and make it work for everyone.
Then repeat for the next one.
Small, steady fixes add up quickly.
Accessibility is central for schools and edtech; our guide to AI in education covers tools for teachers and pupils.
How UIUX Studio helps
We run accessibility audits that combine automated tools, AI-assisted triage and manual testing, and fix issues at the design-system level. Talk to our product design team.
Frequently asked questions
When did the European Accessibility Act start applying?
On 28 June 2025.
Does the EAA apply to non-EU companies?
It can, if they sell covered products or services to EU consumers.
Can AI make my site compliant?
AI helps find and fix issues, but compliance needs manual testing and judgement.
Which WCAG level should I target?
Level AA is the usual target.
What is EN 301 549?
The European standard for ICT accessibility, which incorporates WCAG criteria.