Making chatbots accessible: A guide to enhance usability for users with disabilities

Enhance chatbot usability for users with disabilities: Learn strategies for accessibility, including visual, cognitive, and compliance aspects.

Previously, we discussed the Urgent Need for Chatbots to be Accessible. In this piece, we are taking a deep dive into actions one can take to enhance the accessibility of chatbots for users with disabilities. In order to meet that end, programmers can implement several specific strategies that address visual, auditory, motor, and cognitive impairments. This guide offers detailed instructions, practical code examples, and tool recommendations to make chatbots more accessible.

Let’s get started.

Visual accessibility

• Screen reader compatibility: Ensure chatbots are compatible with screen readers by using semantic HTML and ARIA roles.
• Alternative text for images: Provide descriptive alt text for all images, icons, and emojis within the chatbot interface.
• High contrast and customizable visuals: Implement high contrast color schemes and allow users to customize colors. Ensure text and background colors meet WCAG minimum contrast ratios.

Auditory accessibility

• Text-based alternatives: Offer text-based alternatives for any audio content, including sound notifications and voice responses. Use Web Speech API for integrating text-to-speech and speech-to-text functionalities.
• Visual notifications: Use visual indicators for alerts and notifications to ensure that users who are deaf or hard of hearing are aware of new messages or status changes.

Motor accessibility

• Keyboard navigation: Ensure the chatbot interface can be fully navigated using a keyboard. Implement focus management to guide users through interactive elements.
• Voice control support: Integrate voice recognition technology to allow users with motor impairments to interact with the chatbot through voice commands.

Cognitive accessibility

• Simplified language and clear instructions: Use plain language and provide clear, concise instructions and feedback. Tool: Use language simplification software like Hemingway Editor to ensure clarity.
• Consistent and predictable interactions: Design chatbot interactions to be consistent and predictable, avoiding sudden changes that could confuse users with cognitive disabilities.

Universal design features

• Multiple interaction modes: Offer various modes of interaction, such as text, touch, and voice, to accommodate different user preferences and needs.
• Customizable settings: Allow users to customize settings such as text size, fonts, and speed of voice responses.

Testing and feedback

• Inclusive testing: Regularly test the chatbot with a diverse group of users, including those with disabilities. Best Practice: Involve users with disabilities in the testing phase to gather authentic feedback.
• Use of accessibility evaluation tools: Employ automated tools and manual testing methods to assess the chatbot’s compliance with accessibility standards like WCAG. You can also go further and use SiteLint, which will audit the chatbot automatically whenever a user uses the application, or the chatbot itself will generate various behaviors automatically.

Compliance and ethical considerations

• Adherence to legal standards: Ensure the chatbot meets legal requirements for accessibility under laws like the ADA.
• Ethical design practices: Commit to ethical design practices that prioritize inclusivity and the dignity of all users.

Technical advises

The blog post Web Chat accessibility considerations on Craig Abbott’s website discusses the importance of accessibility considerations for web chat features, emphasizing the need to ensure that chat interfaces are usable by individuals with various disabilities.

The post outlines several key points:

• Accessibility features: Web chat requires the use of ARIA (Accessible Rich Internet Applications) features to ensure real-time information and notifications are accessible to users with disabilities. This includes considering the Web Content Accessibility Guidelines (WCAG) to provide the best experience for screen reader users.
• Chat interface variations: The accessibility impact varies depending on whether the chat feature is a full-page application or a smaller window. This affects how users interact with the interface, especially those using keyboards, screen readers, or dictation software.
• Announcing new messages: The method of announcing new messages to users is crucial. Options include reading out messages in full, only when the chat window or feature has focus, or using secondary indicators like sounds or notification icons. The persistence of chat and notifications across multiple browser windows or tabs is also considered.
• Complexity of chat feature: Balancing the amount of information read out to avoid overwhelming users while ensuring vital information is not missed is challenging. The post suggests making the chat easy to find and use, with features available to multiple senses to enhance accessibility.
• Semantic markup and ARIA attributes: Using role="log" with an accessible name for chat windows helps semantically mark up the chat feature, making it more accessible to screen readers. Testing with assistive technologies and providing alternatives for communication (e.g., phone calls or sign language interpreters) is recommended.
• Accessible names for buttons: Providing an accessible name for buttons that open the chat feature, such as Chat with customer support, ensures that screen readers can provide context. This can be achieved through physical text, alt text on images, or aria-label attributes.
• Visual relationships and visually hidden content: Marking up visual relationships between content (e.g., distinguishing between messages from different users) and using visually hidden content to provide additional context for screen readers are important for accessibility.
• Value and sensory characteristics: Maintaining the current state of controls (e.g., whether a chat window is expanded or not) and ensuring content is available to all senses (e.g., using visual, auditory, and haptic feedback for new messages) are critical for accessibility.
• Status messages and log role: Using role="log" for chat windows allows for the creation of accessible chat logs that screen readers can read out. This is particularly useful for updating users about the chat status (e.g., new messages, chat has ended).
• WCAG criteria: The post highlights several WCAG criteria that are essential to consider for chat features, including those related to information and relationships (1.3.1 Info and Relationships), meaningful sequence (1.3.2 Meaningful Sequence), sensory characteristics (1.3.3 Sensory Characteristics), orientation (1.3.4 Orientation), contrast (1.4.3 Contrast (Minimum)), reflow (1.4.10 Reflow), text alternative (1.4.11 Non-text Contrast), keyboard navigation (2.1.1 Keyboard), no keyboard trap (2.1.2 No Keyboard Trap), adjustable timing (2.2.1 Timing Adjustable), controlling features (2.2.2 Pause, Stop, Hide), managing focus (2.4.11 Focus Not Obscured (Minimum)), using appropriate name, role and values for correctly identify elements (4.1.2 Name, Role, Value), and manage statuses (4.1.3 Status Message), to ensure that chat is accessible to all users.

General recommendations for choosing an accessible chatbot platform

While specific platforms were not recommended here, we recommend to prioritize customizability, compliance with accessibility standards, and support for assistive technologies when selecting a chatbot platform. Additionally, conducting thorough testing with assistive technologies and considering the accessibility of all content types are crucial steps in developing an accessible chatbot.

Customizability

Look for platforms that allow extensive customization of the chatbot’s UI and functionality. This is crucial for implementing accessibility features that might not be available out-of-the-box.

Compliance with WCAG standards

Ensure the platform supports or can be configured to comply with the latest Web Content Accessibility Guidelines (WCAG), which are the internationally recognized standards for web accessibility.

Testing capabilities

Choose a platform that facilitates testing with various assistive technologies, such as screen readers (JAWS, NVDA, VoiceOver, TalkBack). This ensures that the chatbot can be accessed and used by individuals with disabilities.

Support for rich media accessibility

The platform should support adding alternative text for images, captions for videos, and transcripts for audio content. This ensures that all types of content are accessible to users with different abilities.

Keyboard navigation

The platform should inherently support or allow for the implementation of full keyboard navigation, enabling users who cannot use a mouse to interact with the chatbot.

Ensure that the focus indicator is also available. Focus indicators are crucial for accessibility because they provide a visual cue indicating which element is currently selected or active, especially for users navigating primarily through a keyboard or assistive technologies.

Final words

Chat is a complex feature and probably one of the most challenging in terms of balancing user experience and technologies. In addition, there are technological challenges such as the number of messages sent, synchronization, adaptation to devices, online / offline work, security, etc.

In the end, don’t forget to always test with real users, including those with disabilities.