liveBook · Manning

Multimodal interaction

As explained in Effective Conversational AI

Multimodal support is the provision of conversational assistance through two or more interaction modalities, such as voice, text, or visual content.

Multimodal interaction combines channels such as text, voice, images, and video. It can make an assistant more versatile, but combining channels adds design complexity. Choose modalities around user needs and the strengths of each channel. section 9.2 section 9.2.1

A spoken question with visual results

A user asks a question aloud; the assistant displays relevant photos and replies in both text and voice. A voice bot might also offer a video tutorial or other visual aid. section 9.2.1 section 9.2.4

Table 9.3 Modality overview
Text modality
Visual modality
Voice modality
Multimodal interactions
Description
Communication through written text
Communication using text with additional visual elements
Interaction via spoken language
Combines text, voice, and visual elements
Strengths
Precise communication.
Easy to reference.
Enhances comprehension with visuals through an engaging interface
Hands-free operation.
Natural conversational flow.
Richer user experience.
Versatile interaction.
Challenges
Typing can be slow.
May lack emotional tone.
Visuals may distract or be tuned out.
Requires a well-designed UI.
Requires clear speech.
Can misinterpret accents.
More complex to design and implement
When to use
In situations where typing is convenient
Scenarios needing visual aid for better understanding
Situations requiring hands-free interaction
Scenarios needing a combination of modalities
Effect on user engagement
High engagement in text-centric contexts
Improved engagement with visual aids
Increased engagement due to natural interaction
Enhanced engagement through diverse interaction modes
Evaluation techniques
User feedback surveys.
Usability testing.
User experience testing.
Visual effectiveness analysis.
Speech recognition accuracy tests.
User feedback.
A/B testing.
Multimodal usability assessments.

Compare text, visual, voice, and multimodal interactions by their strengths, challenges, and suitable contexts. section 9.2.1

How Effective Conversational AI covers it

How deep it goes
Compares text, visual, voice, and multimodal channels by strengths, challenges, use cases, and engagement.
Examples
Shows a spoken query answered with photos, text, and voice, and voice-bot guidance supplemented by visual aids.
What you can do after
Apply modality comparisons to tailor assistant interactions to user needs and channel strengths.
How it connects
Connects text, voice, visual interfaces, interaction design, and user-engagement evaluation.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Effective Conversational AI. The section links point to the parts of the book it draws on.