Make your AI integrations accessible to everyone

Many of our clients are looking for expert advice about how to esnure that AI-powered services such as chatbots are delivering the accessibility, including:

journey-based accessibility testing;
inclusive user-research;
bespoke expert reviews and consultancy.

Chatbots, AI-powered search, enterprise solutions (such as Microsoft Copilot) and agentic experiences are increasingly common features in websites and apps. Like any other digital feature, AI integrations need to be accessible to disabled users.

AI has the potential to make digital services more accessible. It can also create new barriers if accessibility isn’t considered during integration. Generated content can confuse screen readers, voice input can fail for users with atypical speech, and AI-driven decisions can disadvantage users whose access needs were not represented in training data or testing.

AbilityNet has supported organisations with their accessibility goals for over 25 years. Our expert consultants apply the same audit, user research and consultancy methods we use across web and digital services to your AI features, so you can have confidence you are reaching every customer on every platform.

Contact our accessibility experts

Our AI accessibility services

Targeted AI component and high-risk journey audits

A full audit isn’t always necessary when you are launching a new AI feature. Our consultants can audit just the AI components, such as chat interfaces, generative UI (user interface) or agentic flows, alongside the high-risk customer or colleague journeys that depend on them. This is a focused way to identify accessibility issues in the AI feature without re-testing parts of your product that already have recent audit coverage.

Full WCAG audits and VPATs

For AI integrations that span an entire platform, we offer a full audit against the success criteria of the Web Content Accessibility Guidelines (WCAG) 2.2. Many of our expert consultants are Certified Professionals in Web Accessibility (CPWA) with the International Association of Accessibility Professionals (IAAP). They will outline non-conformant areas, such as insufficient colour contrast, missing alternative text for images, and elements that are not accessible to assistive technology users, and identify areas for improvement.

Inclusive user research with diverse participants

AI features behave differently from one user to the next, which makes automated testing and audits insufficient on their own. User testing can identify up to 45% more accessibility issues than an automated accessibility audit alone, and that gap is often wider for AI features.

Our expert consultants will design a research protocol around the access needs your AI feature is most likely to affect, recruit a diverse panel of participants from our UK network of disabled user testers, run the sessions and provide a detailed report of findings. The panel typically includes users of screen readers, magnification tools, voice recognition software and switch devices, alongside neurodivergent participants.

Discuss an audit or research project

Bespoke AI consultancy for development teams

As well as auditing your finished product, we can work with your engineering, platform and QA (quality assurance) teams to build accessibility into the software development lifecycle. Our consultants offer ad-hoc, bespoke support on:

Accessibility guardrails for AI-assisted development. AI coding tools such as GitHub Copilot, Cursor and Claude Code often produce code with accessibility issues, including missing form labels, low colour contrast, ARIA misuse and broken focus management. We help you put linting rules, PR checks and agent instructions in place that catch these issues at the point the code is written.
Prompt engineering for accessible code generation. The same AI model will produce different code depending on how it is prompted. We work with your platform team to embed accessibility into the system prompts, code generation templates and project-level instructions your developers use, so accessible output is the default.
Automation test coverage for accessibility. Automated tests will only catch a portion of accessibility issues, but the right setup catches many of the ones that matter. We review your CI/CD pipelines and testing frameworks, recommend the right tools and help you write targeted assertions for your component library and AI features, so accessibility regressions are caught as the product changes.

Talk to a senior consultant

Why choose AbilityNet?

AbilityNet is a UK charity specialising in digital accessibility and usability. We have supported organisations with their accessibility goals for over 25 years, with a range of services that includes accessibility audits, user testing, accessibility consultancy and training.

One of the largest teams of specialist consultants in the UK, many of whom are Certified Professionals in Web Accessibility (CPWA) with the IAAP
Founding member of the IAAP and host of TechShare Pro, Europe’s largest annual accessibility and inclusion gathering
A network of disabled user testers across the UK who provide an authentic account of the issues they encounter
Many of our consultants have lived experience of disability and neurodivergence, bringing deep insight alongside their professional backgrounds in psychology, education, UX (user experience) and engineering
Trusted by organisations across financial services, retail, education, government and technology

Talk to us about your AI accessibility project

We offer tailored accessibility and usability support to clients from all sectors. Talk to us about your project and our experts will advise on a bespoke accessibility strategy to meet your requirements.

You can contact us by:

Using the form below
Calling our Digital Accessibility Services team on +44 (0)1926 562 672
Sending an email to sales@abilitynet.org.uk

Contact our accessibility experts

If the form isn’t visible, please email us at sales@abilitynet.org.uk and we’ll respond promptly to help with your enquiry.

Fields marked with an asterisk (*) are required.

Frequently asked questions

Does WCAG apply to AI features?

Yes. WCAG applies to web content and digital services regardless of how the content is generated, so AI-powered chat, generated UI, AI-assisted forms and agentic flows need to meet the same accessibility success criteria as the rest of your product. WCAG 2.2 Level AA is the standard most organisations work to. It is referenced by the European Accessibility Act, EN 301 549 and the UK public sector accessibility regulations.

What is the difference between an AI accessibility audit and a standard WCAG audit?

The success criteria are the same. The difference is in scope and where we focus testing time. AI features tend to have failure modes that don’t always show up in a static page-by-page audit, including streaming responses, dynamic UI, voice and multimodal interaction, and agentic chains. An AI accessibility audit puts those features at the centre of the testing and looks at how the experience holds up for users on a screen reader, magnifier, voice input or switch device, and for users with cognitive access needs.

Can you audit only the AI parts of our product?

Yes. Our targeted component and high-risk journey audits focus testing on the AI features or journeys that have changed, without re-auditing parts of your product that already have recent coverage.

Does the EU AI Act require accessibility?

Yes, for high-risk AI systems. Article 16 of the EU AI Act requires providers of high-risk AI to comply with the accessibility obligations set out in the European Accessibility Act and the Web Accessibility Directive. Many AI systems used in customer-facing services or in decision-making for hiring, finance, education and essential services fall into the high-risk category.

How do you do user research on AI features?

Our expert consultants will design a research protocol around the specific access needs your AI feature is most likely to affect. We then recruit a diverse panel of disabled participants from our UK network, run moderated sessions and provide a detailed report of findings with prioritised recommendations. For AI features we typically include screen reader users, voice input users, users with atypical speech, magnifier users and neurodivergent participants, depending on the journey being tested.

How much does an AI accessibility audit cost?

Many factors influence cost, including the size of the product and whether the AI features are publicly available or behind authentication.