2018年智能语音用户体验报告.pdf

返回 相关 举报
2018年智能语音用户体验报告.pdf_第1页
第1页 / 共45页
2018年智能语音用户体验报告.pdf_第2页
第2页 / 共45页
2018年智能语音用户体验报告.pdf_第3页
第3页 / 共45页
2018年智能语音用户体验报告.pdf_第4页
第4页 / 共45页
2018年智能语音用户体验报告.pdf_第5页
第5页 / 共45页
亲,该文档总共45页,到这儿已超出免费预览范围,如果喜欢就下载吧!
资源描述
VOICE UXBEST PRACTICESLESSONS FROM 17 EXPERTSGIVING VOICE TO A REVOLUTIONTable of Contents About VoicebotVoicebot produces the leading online publication, newsletter and podcast focused on the voice and AI industries. Thousands of entrepreneurs, developers, investors, analysts and other industry leaders look to Voicebot each week for the latest news, data, analysis and insights defining the trajectory of the next great computing platform. At Voicebot, we give voice to a revolution.About ApplauseGIVING VOICE TO A REVOLUTIONApplause is the worldwide leader in crowd-sourced digital quality testing. With 300,000+ testers available on-demand around the globe, Applause provides brands with a full suite of testing and feedback capabilities. This approach drastically improves testing coverage, eliminates the limitations of offshoring and traditional QA labs, and speeds time-to-market for voice, websites, mobile apps, IoT, and in-store experiences.Thousands of leading companies including Ford, Fox, Google, and Dow Jones rely on Applause as a best practice to deliver high quality digital experiences that customers love. Learn more at applause.INTRODUCTION / 3FOUNDATIONAL CONCEPTSBUILDING FOR USERS / 7BUILDING FOR CONVERSATION / 9HOW VOICE DIFFERS FROM OTHER INTERACTION METHODSHOW VOICE CONVERSATION IS DIFFERENT FROM IVR / 12HOW CHATBOTS AND VOICE ARE DIFFERENT / 15HOW VOICE AND VISUAL ARE DIFFERENT / 18ADVANCED CONCEPTSMULTIMODAL DESIGN / 22MULTI-CONTEXT AND MULTI-CHANNEL DESIGN / 25THE POWER OF TESTING AND TUNING / 27BRANDING IN VOICE / 32BEST PRACTICES, COMMON MISTAKES AND RESOURCESBEST PRACTICES AND COMMON MISTAKES TO TO AVOID / 37CONTRIBUTORS / 43ADDITIONAL RESOURCES / 44 VOICEBOT.AI - All Rights Reserved 2018Voice is changing the way we interact with our personal deviceseven within our larger environment. From telling Alexa to set a rice timer for 10 minutes to asking Google Assistant to make calls on your behalf, interactions with voice AI are becoming increasingly common as technological capabilities evolve. Voice isnt exactly new even in the technology space. Speech recognition has been quite capable for at least 20 years and Siri has been helping us send text messages since 2011. PEOPLE ARETALKING VOICEBOT.AI - All Rights Reserved 2018PAGE 3What is different is the rapid expansion of use cases and access to voice assistants. Whereas voice technology was once synonymous with the dreaded interactive voice response (IVR) phone trees that were designed to help curtail customer service costs while wasting time for consumers, voice assistants have changed perceptions and accelerated frequency of use seemingly overnight. Voice is everywhere. But, are we implementing these new solutions properly?Teaching Computers to Understand UsUsing voice to access digital content and execute tasks is very different than click, touch and swipe. Previously, we had to learn how to speak a computers language by either employing manual inputs or occasionally adhering to a narrow set of speech interactions that digital devices were trained to understand. We arent learning how to talk to computers anymore. We are teaching computers to talk to and understand us. The shift is profound. It places a larger burden on voice app developers to deliver a seamless voice user experience in order to lessen the burden on the user. However, you are not alone in confronting these challenges. Other people have accumulated decades worth of expertise in designing voice user experiences while some have come to the discipline recently but are already offering many new insights. Over the past year, the Voicebot Podcast has conducted over fifty hours of interviews and many of those discussions centered on voice user experience design. Some of the guests include authors and long-time voice UX design practitioners such as Cathy Pearl and Lisa Falkson. Other guests such as Tim McElreath and Mark Webster brought visual design expertise to the problem and learned how to modify it for voice. These designers by training have been complemented by linguists and technologists such as Tobias Geobel, Noelle LaChartie and Jan Knig that introduced an entirely new perspective. VOICEBOT.AI - All Rights Reserved 2018PAGE 4In all, we have amassed over 100 insights, tips and techniques from 17 voice UX experts in the Voice UX Best Practices. The eBook was created for designers, technologists, marketers and managers, especially those focused today on building for Amazon Alexa, Google Assistant and other emerging voice assistants. It provides a quick overview of the most common challenges, pitfalls and opportunities associated with voice UX. And, it will hopefully become a trusted resource to share with your teams when they need to get a quick overview of key considerations associated with the discipline. The eBook is organized in four sections with 12 subsections listed on the right side. We hope everyone finds the information helpful and easily accessible for the experienced practitioner and novice alike. There will be a companion podcast where you can hear from the experts themselves in their own voices as a supplement to these materials, but both will stand on their own. If you have questions about the content or topics that were not covered, feel free to connect with Voicebot on Twitter or email us at infovoicebot.ai.EBOOK SECTIONSFOUNDATIONAL CONCEPTSBuilding for UsersBuilding for ConversationHOW VOICE DIFFERS FROM OTHER INTERACTION METHODSHow Voice Conversation is Different from IVRHow Chatbots and Voice Are DifferentHow Voice and Visual are DifferentADVANCED CONCEPTSMultimodal DesignMulti-Context and Multi-Channel DesignThe Power of Testing and TuningBranding in Voice BEST PRACTICES, COMMON MISTAKES dont over-promise.Acknowledge your users.Make sure the use case guides your decisions EXPERT EXAMPLECATHY PEARL“If you want a date from somebodyyoure doing a travel appand you say, “when do you want to travel?” You might have all your dates plugged in, and then someone says, “I want to leave next Tuesday evening.” Maybe you did not expect that. Modifying prompts is a simple solution: if you prime someone by saying “what date do you want to leave?” they are much more likely to give you an actual date.”1.2.3.4.PAGE 8 VOICEBOT.AI - All Rights Reserved 2018GET CONVERSATIONALTIM MCELREATH, Discovery Communications“Dont expect that people talk the way you talk, or converse the way you converse. It is crucial to assume that theres going to be a diversity in the way people converse even within a single language.” “One of the things I learned very quickly is in order to design an effective voice experience you have to listen to the response a lot, over a long period of time. You dont know how frustrating a response can get until youve heard it 50, 100 times.” “If you want to design an experience that people start using visually, and come back to over and over again, you have to design it so that they are going to want to sit through a response to the end without getting frustrated.”When building for conversation, dont assume you know how people talk. Not everyone relays information and conveys meaning in the same way. Even when you think a particular response to a prompt is obvious, that is rarely the case. This is the crux of how conversation is a fundamentally different way to interact with technology than where we started. “Design for how people actually talk, not how you want them to talk. You can modify prompts to cater to this, as the prompts themselves are key in getting back the kind of information you want,” Cathy Pearl, Author of Designing Voice User InterfacesAccommodating the Computers LimitationsHistorically, the communication and understanding capabilities of computers was severely limited. Computers only knew their own language and input mechanisms were devised to enable humans to input instructions. Initially that meant communicating directly in the computers language by typing out code or commands (or using punch cards created on special typewriters) on a screen. Building for ConversationPAGE 9 VOICEBOT.AI - All Rights Reserved 2018This interaction method was displaced with the introduction of new user interfaces (UI) that abstracted the communication method and translated more human-friendly input methods into computer language in the background. In turn, these UIs translated the computer response into formats more easily consumed by humans. However, these were all programmatic interfaces. There were a predefined set of interactions with no way to go beyond them. Menus and buttons made it easier to interact with computers while clearly proscribing the limitations of our interactions. If the option was not on the screen (i.e. the UI) then it wasnt an option. This reduced the incidence of errors, but forced users to learn the systems in order to use them. Humans by necessity adjusted their communication methods to account for the limitations of computers. Beyond Programmatic InterfacesThe advent and advancement of Natural Language Processing (NLP) with its Automated Speech Recognition (ASR) and Natural Language Understanding (NLU) capabilities introduced the option for non-programmatic inputs. It has flipped the dynamic enabling humans to communicate in their preferred method and computers to accommodate, adjust and learn. Its an extraordinary change and UX designers are vested with the responsibility to help computers succeed in the new dynamic. This is the context behind the UX designers requirement to design for conversation. It is not just about helping the user, but also helping the computer, the voice assistant, be successful in its interactions. So, designing for conversation presents two challenges. First, you must design for both the user and computer. Second, you must design for interactions without the boundaries many of us became accustomed to in programmatic interfaces. As you move into conversational design, experts say critical success criteria include not assuming you know how users will speak and evaluating actual conversations to really understand what is working and what is not.To Build for Better ConversationDesign for how people actually talk. Assume diversity in the ways people talk and converse.Use carefully-constructed prompts to get back the information you want. Dont bury your prompt in the middle of your response. Put the prompt at the end so it is the last item the user is reacting to.1.2.3. VOICEBOT.AI - All Rights Reserved 2018PAGE 10
展开阅读全文
相关资源
相关搜索
资源标签

copyright@ 2017-2022 报告吧 版权所有
经营许可证编号:宁ICP备17002310号 | 增值电信业务经营许可证编号:宁B2-20200018  | 宁公网安备64010602000642