VOICE Summit 2019 | VED Talks | Dylan Zwick | July 23
summary
Voice interfaces require handling unconstrained speech, sequential and ephemeral output, and making speech the easiest input; Pulse Labs tests show three concrete design rules: adaptable (map synonyms like “Johnnie Walker” to whiskey rather than fail), contextual (remember scores and recent actions so you don't reply “too tough” after a correct answer), and available (avoid deep menus so a “find Wrinkle in Time” utterance performs the lookup immediately). Pulse Labs tests 150+ brands and ran Alexa managed partner UX last year.
AI-written summary · about
chapters (14)
- 0:00Intro
- 0:30Conversational Design
- 2:28Voice
- 2:57Adaptability
- 4:34Example
- 5:46Be contextual
- 6:19Individualization
- 6:43Random
- 7:47Jeopardy
- 8:14Availability
- 8:34Nested menus
- 9:04Wide toplevel UI
- 9:32Ticket ordering example
- 10:23Food reordering example
related talks
19mAlexa, show me a good voice UI
16mWeak signals of a vocal revolution in UIs
44mDelivering On The Promise Of Voice With Alexa
21mMulti modal interfaces and the future of voice
24mVoice design best practices
02mInterview with Joaquín Engelmo, Speaker at Commit Conf 2019
50mDesigning Experiences For Voice Interactions
20mVoice Interfaces in Everyday Life