accessiBe is adding a conversational AI layer to its accessibility widget, giving website visitors a new way to request accessibility adjustments, ask questions and find information without having to work through a traditional settings menu.
The company announced the launch of its AI accessibility assistant on August 10, saying the new feature is intended to make the interaction between a visitor and an accessible website more direct.
Instead of figuring out which accessibility setting they need, visitors can describe the problem in ordinary language.
They can say, “the text is too small.” They can ask, “take me to the footer.” Or they can simply ask, “what’s this page about?”
The assistant then interprets the request and takes an appropriate action.
A Different Way to Interact With Website Accessibility
Website accessibility tools have traditionally relied on menus, buttons and predefined settings. A visitor may need to identify a particular adjustment and then locate the corresponding control.
That approach can work. But it assumes the visitor already knows what setting to select.
accessiBe’s new AI accessibility assistant takes a different approach. The visitor describes the desired result rather than identifying the technical adjustment.
That distinction may sound small, but it could be important.
A person does not necessarily think, “I need to activate a specific text-sizing profile.” They are more likely to think, “I can’t read this text. Make it bigger.”
The assistant is intended to bridge that gap.
Robert Lopez, CEO of accessiBe, said the accessibility adjustments were already available through the company’s technology. Conversational AI gives visitors a more natural way to access them.
That is probably the most interesting part of the announcement. The company is not presenting conversational AI as a replacement for its accessibility functionality. Instead, it is using AI as a new interface for capabilities that already exist.
Visitors Can Describe the Problem in Their Own Words
The AI accessibility assistant is built around natural-language interaction.
Natural language means that users can communicate in ordinary words rather than having to learn a particular command structure or understand the terminology used by accessibility software.
According to accessiBe, the assistant can perform several functions based on what a visitor asks it to do.
- Apply the appropriate accessibility adjustment or profile.
- Answer questions about the content of the current webpage.
- Provide a plain-language summary of the page.
- Guide visitors to specific elements on the page.
- Shift focus to the requested element.
That gives the assistant a role that extends beyond visual accessibility controls.
For example, a visitor could ask what a page is about without having to read through the entire page first. Someone looking for a specific part of a webpage could ask the assistant to take them there. A visitor struggling with text size could request a change without knowing which control makes that adjustment.
The common thread is that the visitor describes the outcome they want.
The technology handles the translation from that request to the corresponding action.
AI Could Make Accessibility Tools Easier to Understand
There is a broader usability issue here that deserves attention.
Accessibility settings can become complicated. Websites may offer controls for text size, contrast, color, animation, focus, navigation and other aspects of the user experience.
For someone who understands those controls, a settings panel can be useful.
For someone who does not, it can become another interface that needs to be learned.
Conversational AI potentially removes some of that friction.
The visitor does not have to know the name of an accessibility feature. They can explain the problem instead.
That could be especially useful for people who encounter a website with an accessibility need but are unfamiliar with the terminology commonly used by accessibility tools.
It also changes the interaction from “find the correct setting” to “tell the system what you need.”
The Existing Settings Menu Is Staying
accessiBe is not eliminating the existing accessWidget settings menu.
The company says the menu will remain available for visitors who prefer to browse the available adjustments and choose them directly.
That is an important detail.
Conversational interfaces are not necessarily better for everyone. Some users prefer visible controls. Others may want to see exactly which settings are active. A traditional interface also provides a predictable way to interact with accessibility features without relying on an AI interpretation.
Keeping both options gives visitors a choice.
They can use the conversational assistant, or they can continue using the existing settings interface.
accessiBe Says People With Disabilities Helped Build It
accessiBe also emphasized how the AI assistant was developed.
The company says it worked with its Inclusive Product Advisory Board and accessLabs throughout development. Both groups include people with disabilities who provide input on accessiBe’s products.
This is an important part of an accessibility product launch because there is a meaningful difference between building a feature for people with disabilities and building one with their participation.
Technology teams can make reasonable assumptions about how people will use a product. Those assumptions do not always survive contact with actual users.
Feedback from people who use accessibility technology can expose problems that may otherwise be missed during development.
Lopez said the groups helped shape the assistant from the beginning and will continue to influence its development.
That ongoing involvement may become particularly important as the company collects more information about how people interact with the conversational interface.
The Assistant Is Part of a Larger Accessibility Platform
The new AI assistant is one component of accessiBe’s broader accessibility platform.
The company also offers accessFlow, which provides developer tools for code-level remediation. Code-level remediation means making changes directly to a website’s underlying code so accessibility issues can be addressed at the source.
accessiBe also offers accessServices, which provides expert-led manual testing, audits and litigation support.
That combination gives accessiBe three distinct technology and service layers.
AI automation can address certain accessibility needs through the user interface. Developer tools can help organizations address issues in their underlying code. Expert services can provide human testing, assessment and support.
Those functions address different parts of the accessibility process.
That matters because no single interface or automated system can reasonably be expected to solve every accessibility issue on every website.
Where Conversational AI Fits Into Web Accessibility
The addition of conversational AI reflects a larger trend across software.
Interfaces are increasingly moving from menus and commands toward natural-language interaction. Instead of searching through a series of controls, users can describe what they want.
Accessibility may be an especially interesting application for that model.
A traditional interface generally requires the user to understand the structure of the software. A conversational interface can potentially allow the software to understand the user instead.
That does not mean AI will always interpret a request correctly. It also does not remove the need for established accessibility standards, testing or properly built websites.
It does, however, offer another way for people to interact with digital content.
And that distinction is worth watching.
Automation Does Not Replace the Need for Accessible Websites
There is also an important caveat for website owners.
An accessibility widget should not be viewed as a substitute for building an accessible website in the first place.
Web accessibility involves the underlying structure and functionality of a website. That includes semantic HTML, keyboard navigation, meaningful labels, alternative text, form accessibility, focus management, color contrast and other technical considerations.
Automated tools can play a role in that process. Developer remediation and manual testing can play another.
The strongest accessibility strategy generally involves addressing accessibility throughout the development and maintenance process rather than treating it as a final layer added after a website is built.
That is one reason accessiBe’s positioning of automation, developer tools and expert services as complementary components is noteworthy.
accessiBe Starts the Rollout With Growth Plans and Higher
The AI accessibility assistant is live for accessWidget customers on Growth plans and higher, according to accessiBe.
The company says it plans to expand availability to additional customers and partners through a phased rollout.
That rollout should provide accessiBe with an opportunity to see how visitors actually use conversational accessibility features at scale.
The most useful measurements may eventually extend beyond the number of people who activate the assistant. How accurately does it interpret requests? Which types of requests are most common? How often does a visitor need to rephrase a request? Which accessibility adjustments are used most frequently?
Those questions could help determine whether conversational AI becomes a meaningful improvement to accessibility or simply another interface option.
A More Natural Conversation With a Website
accessiBe says more than 85,000 websites worldwide use its platform. The company has also worked with organizations including United Cerebral Palsy, the Special Olympics USA Games and the Parkinson’s Foundation.
The new AI assistant gives those accessibility efforts another technological layer.
The concept itself is straightforward. Instead of asking visitors to understand the controls, the website can ask them to explain what they need.
That is a subtle change in interface design, but potentially a significant one for accessibility.
There will be questions about accuracy, privacy, reliability and how conversational AI handles ambiguous requests. Those issues will need to be evaluated as the feature reaches more websites and users.
Still, the basic idea is compelling. Accessibility technology should make digital experiences easier to use, not require users to learn another technical system before they can get the help they need.
By putting a conversational interface directly into accessWidget, accessiBe is betting that sometimes the simplest accessibility control is a person being able to say what they need in their own words.