Recommendation 4.3.2:
Support speech-to-text (STT) user input and text-to-speech (TTS) chatbot output
Benefits Users
Visual | Cognitive | Physical
Relevant W3C Guidance
Phase 1: Gather & Organize
[none]
Phase 2: Design & Implement
Design Question
-
Is built-in speech-to-text or text-to-speech worth the investment for your chatbot?
- Built-in speech-to-text and text-to-speech functionality can be trained for the unique words or acronyms in your application and ensures that your content is communicated exactly how you would like. It also supports situationally disabled users (e.g., a user driving a car is temporarily unable to view the screen) who may not have assistive technology available. On the other hand, users with assistive technology have (ostensibly) configured to their devices to needs and preferences. In addition, users’ devices are their own responsibility (and cost) to update as technology advances.
Tip
- If your chatbot offers built-in text-to-speech, make sure the user knows how to easily turn this feature on and off.
Phase 3: Test & Evaluate
Self Check
-
The chatbot either supports speech-to-text or behaves equivalently when a third-party speech-to-text device or software is used as an input.
-
The chatbot either offers text-to-speech or communicates equivalently when a third-party text-to-speech device or software is used.
Ask the User
-
Is the chatbot experience and capability similar whether you’re using your own TTS/STT device or a built-in functionality?
- This question is subjective; use a Likert scale. This question should be asked post-study.