Designing Voice User Interfaces: Principles of Conversational Experiences

by Cathy Pearl

Margaret Urban, interaction

Priming refers to the fact that exposing someone to a particular stimulus (such as a word or image) will influence their response

certain number of questions is also a form of priming;

coreference, and is an essential part of communication.

Note

result.

But it should be noted that it doesn't read this out loud.

context-specific help.

needs.

I would try to give them help, as opposed to doing nothing, which is essentially what searching the web for that term is, for the user.

Users need a way to find out just what it is they can really do.

song.

How do you know what they intended?

Prompt lists serve multiple purposes:

Voice User Interface Design.

phrases

I'm assuming this uses fuzzy matching

Conversant Labs,

impaired?

Not true if you're relying on whole or largely on another visual UI.

immersive computing (virtual and augmented reality).

around.

Not necessarily true.

speaking.

Do we have a choice (on iOS)?

barge-in.)

[silence] (after an app prompts the user for input).

Text-to-Speech Personalization

multimodality should ensure that all modalities are available at all times,

Both should always be provided.”

Multimodality should be a user’s choice, and not a requirement.

experience.

What does this mean?

Voice User Interface Design, Cohen, Giangola, and Balogh

Business Insider article “Why Microsoft Doesn’t Want Its Digital Assistant, Cortana, to Sound Too Human,”

responses.[

What does this continuum look like?

product goals, company goals, ethical goals, social goals,

you offer different personas,

Theming for VUI

Ann Thyme-Gobbel during her time as head of UX/UI

Binary Domain

In Verbis Virtus

There Came an Echo

speak

This is not possible on iOS - workaround?

“barge in” (interrupt)

allow the user to continue the conversation, but don’t automatically turn the microphone back on

“The Jack Principles of the Interactive Conversation Interface” in 2002.

Volio

interaction:

Any data on this??

Sensely,

Thyme-Gobbel

(ASR).

Similarities and differences to NLP?

key:

How do we measure these for iOS and Android?!

N-best lists, settable parameters like end-of-speech timeouts, and customized vocabularies.

command.

Not true. I don't consider this barge in, because the system is not speaking, it's displaying other content, via audio, but other content nonetheless.

hotword or magic word

your system does not have barge-in enabled, do not force the user to listen to long lists or long menus.

end-of-speech timeout