Service design is the activity of planning and organizing people, infrastructure, communication, and material components of a service in order to improve its quality and the interaction between service provider and customers. As VUI design becomes part of the large design world, we discuss what the emerging discipline of service design can teach us and which omnichannel applications might provide the most practical “bang for the buck.”

“My New Voice” is a chamber opera in progress. The piece tells the story of Trevor, a character with motor neuron disease, who uses a TTS synthesizer to speak. In this session, the presenters discuss their experience as users of speech technology, show clips from the performance, and generalize learned information to broader industry applicability. Their conclusions demonstrate that such niche applications can provide both practical and deep insights into the tools, capabilities, interfaces, and design constraints of speech technology.

This session discusses how to support and seamlessly augment social, email, direct mail, or live agent interactions with outbound IVR applications. We review several case studies describing how outbound IVR experience delighted customers because they were updated in real time with accurate contextual messaging that was customized on a one-on-one basis for each and every client, and was integrated and in sync with messaging from other channels.

As self-service interactions have shifted to other channels, a strong emphasis on cost containment in operational areas has left a legacy of features intended only to make life easier for a company, not the caller. With creative applications of emerging technologies that can identify and even authenticate callers behind the scenes, is it time to retire the habit of forcing callers through voice-based identification and verification procedures simply to make life easier for an agent?

We will soon be floating in a sea of devices. Most of them will be part of the interconnected Web of mechanisms that will constitute the Internet of Things (IoT). The overwhelming majority of them will talk to each other but not to humans. This presentation focuses on the much smaller set of devices that are articulate. Some articulate mechanisms translate information that bubbles to the surface of the IoT into natural-language messages for humans and they transmit spoken commands given by humans to other IoT devices. One union of speech-enabled and machine-to-machine IoT devices is the smart home. Specifically, this session explores smart-home technology for eldercare. Other connected devices communicate directly between the Internet and humans. This session also focuses on a rapidly growing group of such devices: toys. Combined, these discussions will reveal emerging opportunities for speech and language processing.

Implementing a multilingual self-service solution shared across multiple countries is undoubtedly demanding. The logistical issues of language selection and country-specific elements such as currency, postal codes, phone numbers, addresses, mnemonics, dynamic data, and cultural differences will keep you up at night. This presentation focuses on lessons learned from the implementation of such a solution and includes such topics as complexity concerns around dialectal, cultural, and country-specific peculiarities; planning; and human factors considerations.

Have you ever had trouble using an off-the shelf SML ASR that doesn’t recognize technical words correctly (for example, the commonly used word “ate” is returned instead of the word “eight”)? We describe a post-ASR processing strategy that uses a “soundex” algorithm to identify words that sound like the incorrectly recognized word (“ate”) and select the word that makes sense in the current context (e.g., “eight”). We describe uses for this type of post-ASR processing.

Customer feedback is a powerful tool for improving all aspects of the customer experience. But it has to be done properly to be effective. Attend this session to learn how to bring customer feedback into VUI design, call centers, online self-service, and other customer channels. Hear stories of customer feedback done right (and wrong), and the lessons we can draw from those experiences.

When the SpeechTEK conference began 20 years ago, the speech industry was in its infancy, but in the latter decade, speech has become a fully mature industry. This panel provides a historic perspective of the milestones achieved and discusses the key issues that have shaped the speech industry. Join industry luminaries and big thinkers for a thought-provoking and entertaining look back at where we’ve been, and where we’re heading in the next 10 years.

We’re entering a new era of computing—ubiquitous computing—in which computers are around us, everywhere. As the Internet of Things continues to emerge, speech-enabled interfaces hold the promise of unlocking a new generation of intelligent devices. However, to truly gain mass-market appeal the Interface of Things will need to have near flawless performance. Join us as panelists discuss the opportunities and challenges for speech in the coming years. What are the latest developments in speech recognition and natural language understanding? What developments can we expect in the next 2 years? What barriers remain to be conquered?

The Internet of Things promises to turn many everyday objects into connected platforms to enhance daily living. But how, exactly, do you speak to a refrigerator? How will your sport utility vehicle respond to you? Panelists in this session discuss building new models of interaction and engagement with newly connected objects in our lives. Other topics include barriers to adoption, the importance of predictive assistance, natural language understanding, and how to build trust and relationships.

The Web of Things is growing rapidly and makes possible an incredible variety of new applications. The large variety of connected devices, with widely differing capabilities, would require mastering dozens or even hundreds of different graphical user interfaces. This presentation shows how spoken natural language can streamline the user interface to the Web of Things by allowing users to more directly and more naturally express their intents.

Voice interaction designers have always been out of step with their peers in visual design in one significant way: the lack of field research. Major visual design projects usually include user research conducted outside a research lab, either during requirements gathering or as part of usability testing. The practice never caught on when speech was confined to the IVR, but now that speech technology is being used everywhere, escaping the lab is vital. Panelists discuss how to make this happen.

The adoption of voice technologies by hundreds of millions of users is a positive development, but highlights the fact that speech recognition does not work equally well for some segments of society, notably children, seniors, and accented speakers. These users may become alienated by poorly performing speech technology and cut off from its benefits. This presentation characterizes the extent of the problem, summarize research in applying speaker adaptation and deep learning to address the problem, and makes a case for preventing a discriminatory “speech divide” via more inclusive speech technology.

Elderly people, particularly those with cognitive impairments, often experience difficulty in performing everyday tasks because they are unable to remember the sequence of steps in a task or which steps they have already completed. Recently, tools using speech and language technologies have been developed that prompt elderly patients with reminders or checks and respond to patients’ questions or feedback. Several examples of these applications are presented, along with a discussion of the technologies involved, and future enhancements are suggested.

This session previews the current state of the growing healthcare wearable technology market and provides a glimpse into the near future of intelligent voice and conversational interfaces in consumer healthcare wearable technologies. Current wearable technology is still in its infancy, can be quite complex to use, and has limited capabilities. The next generation will be intelligent voice-enabled devices and go beyond tracking to interacting with and assisting consumers with their healthcare choices, thereby changing behaviors for the better.

Attend this panel for a discussion of the vital role of voice interaction design in a variety of healthcare applications. The healthcare domain presents unique challenges in terms of confidentiality, persuading individuals to change behaviors, and users who may have physical or mental impairments. Healthcare experiences must not only be effective but show sensitivity to users’ anxiety over an upcoming procedure, fear and denial regarding a recent diagnosis, or confusion over a complicated new medication regime.

There are some excellent tools on the market for creating your own virtual assistant; most are intended for corporate customers with large budgets. But there are also freely available open source tools that enable developers with some background in software development to create their own virtual assistants. We examine, show demos and videos, and compare currently available open source tools including the JiboAlive Toolkit, Virtual Human Toolkit, Teneo Studio, CMU’s Jasper for Raspberry Pi, Speaktoit, and Pandorabot.

You can use the newest, fastest, best technology in the world yet not provide the benefits you anticipated. Based on real- world interviews with hundreds of companies and users, from dozens of new technology implementations and original technology innovation research conducted by Theresa Szczurek, this session explores the seven essential factors that increase the probability of successful implementation of visual IVRs and live assistance.

Many enterprises are evaluating a new alternative to a voice-only IVR system—visual IVR. With mobile smartphones nearly ubiquitous (more than 70% of U.S. adults now own one), the visual interface can reach a mass audience. A visual IVR can deliver both cost efficiencies and a better user experience. Will enterprises now focus their limited resources on the visual interface at the expense of speech? What are the pros and cons of visual versus speech-only IVRs? Is there a way to effectively use both types of interfaces in a single solution? These and other questions are addressed during this session and the subsequent panel discussion.

Written text is permanent, while spoken text is transient. Visual IVR offers users a visual representation of their IVR menus that can greatly improve the IVR experience. Can you add visual IVR capability to your existing voice-only IVR infrastructure to leverage your existing IVR call flows, help keep custom- er service cost down, and extend the life of your legacy IVR investments? What are the similarities and differences among the various approaches to visual IVR?