Conversational Design Theory: Conversational Design Principles

Transcript

1. #givevoice @voxable voxable.io Conversational Design Theory: 
 Conversational Design Principles Austin Chat UX | January 2019 Lauren Golembiewski CEO & Co-Founder, Voxable
2. #givevoice @voxable voxable.io Voxable helps companies design and develop conversational interfaces as well as teaches teams strategy and process. Not much Wanna hang out? Hey, what's up? What about drinks tomorrow at 7? 12>00 PM Active Now Henri Liriani Home 100% 12>00 PM Let's You I return space 123 M N B V C X Z L K J H G F D S A P O I U Y T R E W Q Aa
3. #givevoice @voxable voxable.io Listen Automatic Speech Recognition (ASR) Messaging UI Understand Natural Language Understanding (NLU) Process Bot Intelligence Respond Synthesized Voice Messaging UI The Conversational Interaction Loop
4. #givevoice @voxable voxable.io Listen Research Jobs-to-Be-Done Understand Affinity Maps Object Definition Process Scripts Conversation Flows Prototype Respond UI Elements Branding Writing The Conversational Design Process
5. #givevoice @voxable voxable.io Conversational Design Principles
6. #givevoice @voxable voxable.io The UX of Conversation - Intuitive - Social - Collaborative - Empathetic - Emotional - Temporal - Environmental - Unpredictable - Error prone
7. #givevoice @voxable voxable.io Intuitive - Language is the human interface - Works in a multitude of situations - The way business gets done - Machines make us conscious of the effort to communicate
8. #givevoice @voxable voxable.io Design for Intuition - Understand the users - Document users’ mental models - Build a user-centered information architecture
9. #givevoice @voxable voxable.io Social - Conversational interaction builds a relationship with users - Machines change society and the way humans socialize
10. #givevoice @voxable voxable.io Design for Society - Understand users’ needs as well as the societal implications - Think about the bot’s proactiveness or reactiveness - Develop strong branding to drive the bot’s social presence
11. #givevoice @voxable voxable.io Collaborative - Quantity - not too much or too little information - Quality - truthful and evidence-based - Relevance - information that is pertinent - Delivery - clear, brief, and orderly - Polite - give options, don’t impose Grice’s Maxims
12. #givevoice @voxable voxable.io Design for Collaboration - Craft responses that follow Grice’s Maxims - Focus on providing clear prompts to the users - Employ the GUI to provide structure - Wizard-of-Oz prototype and usability test
13. #givevoice @voxable voxable.io Empathetic - Theory of mind - We constantly infer what others might think, feel, and need - This complex context is (mostly) easy for humans but very difficult for machines
14. #givevoice @voxable voxable.io Design for Empathy - Think about what users say vs. what they mean - Perform user research to understand users deeply - Expose yourself to the users’ situation through contextual interviews - Ask users to think out loud
15. #givevoice @voxable voxable.io Emotional - Language has the power to evoke strong emotions - We share emotion just as much as we share information - Consider how users might express emotion with words
16. #givevoice @voxable voxable.io Design for Emotions - Use technology to measure simple emotions - Use research methods to understand emotions - Craft conversation flows and responses based on emotion - Develop the conversational tone for emotions users experience while interacting with with the bot
17. #givevoice @voxable voxable.io Temporal - Information is not persistent - Voice interfaces and users have to interact in real- time - Just-in-time information delivered effectively
18. #givevoice @voxable voxable.io Design for Time and Attentiveness - Optimize everything for brevity and memory - Use simple language; no jargon - Employ audio and visuals that help users pay attention
19. #givevoice @voxable voxable.io Environmental - Conversational interactions happen in new places and with new devices - Our language is influenced by our setting - Language can more difficult or easy depending on the setting
20. #givevoice @voxable voxable.io Design for Environment - Understand the environment and the conditions in which users are interacting with the bot - Perform contextual interviews to observe users in the environment - Vary the conversation flow based on environment, especially for multi-modal bots
21. #givevoice @voxable voxable.io Unpredictable - Conversations constantly shift and evolve - Original goals are abandoned based on the information discovered in the conversation - Bots should allow for “choose your own adventure” style navigation
22. #givevoice @voxable voxable.io Design for Unpredictability - Strive for a choose-your-own adventure style navigation - Map out users’ journeys based on research - Make error messages contextual to users’ goals
23. #givevoice @voxable voxable.io Error Prone - Human conversation is very error prone - Humans have learned how to handle this well, but machines haven’t - Conversational design needs to invest a great deal in error handling
24. #givevoice @voxable voxable.io Design for Error Proneness - Identify error-prone areas of conversation through research - Invest in strong UX writing - Confirm important steps before executing
25. #givevoice @voxable voxable.io The UX of Conversation - Intuitive - Social - Collaborative - Empathetic - Emotional - Temporal - Environmental - Unpredictable - Error prone
26. #givevoice @voxable voxable.io Teaching Machines to Converse - Use the principles to create the design approach - Perform usability studies with real users to inform the approach - Conversational design is really challenging, meticulous, and exciting
27. voxable.io howdy@voxable.io Give your business a voice with Voxable.