Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Strategies for Conversational Product Experiences

UXAustralia
August 30, 2019

Strategies for Conversational Product Experiences

UXAustralia

August 30, 2019
Tweet

More Decks by UXAustralia

Other Decks in Design

Transcript

  1. Conversa)on, the Human Interface • How we exchange meaning and

    desire for action • Engage and fulfill social contracts • Meaning goes deep and broad • Diverse, wide-ranging, and encompasses multiple media and modes • Not just recognition of words and static definitions “The principle of mutual responsibility. The par6cipants in a conversa6on try to establish, roughly by the ini6a6on of each new contribu6on, the mutual belief that the listeners have understood what the speaker meant in the last u=erance to a criterion sufficient for current purposes.” Clark & Wilkes-Gibbs, Referring as a collabora6ve process
  2. Conversa)onal tech • Utterance - what the person says as

    input • Prompt - what the system says to elicit input or give information • Intent - determined from the results of NLP as an indication of what the person wants, either from direct statement or inference • Slot - a variable detail, e.g. dates or cities • Turn - utterance + prompt pair Ok google, give me weather and traffic
  3. Community of home + online life "Bring together familiar things

    to help me do more" Range of voice and touch controls that make sense Strict UI guidelines emphasizing efficiency Extension of loved "interactive"show "Let me play the game just like on TV" Brief, content from daily game, rewarding Game rules and lingo, interruptible
  4. Designing Conversa)on • Anticipating one half of a customer-led collaborative

    context negotiation • Extreme clarity of purpose and function • Exceptional appreciation of the variation of language • Breadth of response to changing needs, wants, and emotions • Planning interactions with capabilities that exceed the complexities of the conversation • Oh, and make it easier, faster, more enjoyable
  5. 5 Strategies for Conversa)on Tech Don't Make Me Think You’re

    Thinking I'm Thinking People Who Need Pseudo-People Love, Love Me Do Yada, Yada, Ya Doin’ It Wrong Yea Though I Talk in the Uncanny Valley of the Shadow of the Death Star
  6. Human constraints • Memory and distraction • Need for predictability,

    naturalness, and familiarity • Ability to be flexible with context • Constant maintenance and repair
  7. Principles and guidelines applied • Clarity about what to do

    and why • Plan for misinterpretation and spontaneous change of direction • Be careful with open-ended prompts while remaining flexible "How may I help you?" vs "What’re you calling about today?"
  8. Coffee is a perceivable object Coffee can be experienced I

    am perceiving and experiencing coffee You are perceiving and experiencing coffee Coffee has been perceived and experienced by both of us before Coffee is eligible for subjective evaluation, and by more than one person We have similar concepts of subjective evaluation We have the ability to express and understand the subjective evaluation of coffee Coffee is variable in its qualities that can be subjectively evaluated Coffee’s variable qualities and their evaluation are of interest to us We have subjectively experienced a range of coffee We have a similar appreciation of coffee…
  9. Principles and guidelines • Brevity - use words effectively •

    Flexible - support variety and spontaneity • Friendly - be comfortable and enjoyable • Focus - give clarity and progress • Speed - make thoughtful use of time • Ease - make the machine do the work • Help - provide context and repair
  10. Neither Good Person Or Machine • Time limits for attention,

    input duration, and lifespan of interaction • Hard-to-discover offerings • Invisible narrowing of domain and limits of abilities • Failure kills the conversation • Systems and apps are not more sophisticated than the complexity of the task and functionality
  11. We Can Know Customers • What does this mean to

    me? How does it benefit me? • What do I think should happen? What’s my goal? What’s required of? • What do I want to know? In what order and structure? Will I want more? • What delights me? What keeps me interested and wanting to come back? • What might I need to learn? What helps me? How do I know I’m learning the right things?
  12. Principles and guidelines applied • Does this prompt sound like

    a human would say it spontaneously? Conscien<ously? If so, why? If not, how could it? • Does the prompt connect easily to all that come before and a@er it? If so, why? If not, what needs to change? • Does the prompt use the same dialogue markers or key words as others nearby in the flow? If not, what needs to change? • Does the prompt need varia<ons of it based on frequency? If so, why and what do those look like?
  13. Language Is Alive • Changing norms and usages • Emergent

    language from various groups • Not laziness, rudeness, or signs of impending apocalypse • When words die, it’s okay • When new words are born, it’s wonderful • Change is a sign of richness, adaption, intelligence “15-year-old users in 2016 wrote more complex posts than users of any age in 2008.” Gretchen McCulloch, Because Internet
  14. Principles and guidelines applied 1.Write what needs to be communicated,

    using the clearest language possible. 2.Replace any “machine speak” with more everyday-human wording. 3.Revise to use words and structures that are simpler. 4.Revise to remove anything that could be interpreted in a way that seems like it’s a machine telling a human what to do. 5.Now revise to relax and create a feel of ease. 6.Now revise to add words of invita<on and welcome. 7.Now revise that to add words allow predic<on and confidence. 8.Now revise that to feel kinder and point to a posi<ve outcome.