You’ve spent countless hours building your IVR system, making sure everything flows. Now it’s time to hire a voiceover talent to bring it to life. Ensuring the voiceover talent understands your system’s flow and its end user is vital to the success of the final recording. In this session, we discuss best practices in hiring, communicating, and working with voiceover talent to create the most effective end-user experience.

The Voice User Interface Designers Toolkit is an open-source project that includes custom Visio stencils, Visual Basic macros, and Python code that supports designing and maintaining a voice user interface. The tools do not generate code but instead help to partially automate the iterative design process. Learn about the strengths and weaknesses of the Toolkit from designers who use it. Learn about the latest enhancements and usage guidelines and offer suggestions for future enhancements.

Google’s project glass, Vuzix glass, and Telephony One are some of the wearable head gear that will bring a new level of user-computer engagement using speech technologies. How will these devices change the use of personal assistants, social media, location relevant data, first person augmented reality, video recording, gaming and simulated sports. What new applications will these devices enable? What new concerns do these devices raise about privacy and security?

Customers hate bad IVR. In fact, a large majority call with the intent of speaking with an agent. Why? Because prior experience says the IVR will be difficult and frustrating. And when an agent is engaged, expenses rise dramatically. Personalization can help alleviate the symptoms, but most technologies cannot leverage an integral component: caller behavior data. Learn how you can leverage caller behavior data in the IVR to adapt the customer experience … with real time analytics.

When considering Kubrick’s masterpiece, 2001: A Space Odyssey, we have surpassed most of our technological predictions. However, when it comes to talking and comprehending, computers are still far from humanlike capabilities. Where are the missing links? What are we doing wrong? On which problems do researchers around the world spend their time? What can we expect in the next few years? This talk, targeted to a heterogeneous audience, addresses these questions.

This session explores why so many natural speech projects fail to deliver the benefits to both customers and the business and what can be done to avoid these pitfalls. Learn how innovative design, coupled with operational changes, can be used to rebalance the call center operational model to get true value from natural speech interfaces. Some of the benefits include getting infrequent callers to the right agent as fast as possible using natural speech, maintaining self-service rates and satisfaction for frequent callers, and increasing overall customer satisfaction.

Studies show that customers are increasingly asking for proactive outbound communication as a component of customer service and self-service. If done right, outbound notifications can help both improve customer satisfaction and reduce costs. Choosing the right means of communication, respecting customer preferences such as contact hours and do-not-call as well as opt-out lists, and providing meaningful and actionable information in outbound IVR and SMS campaigns show tangible ROI. This session includes real-world case studies from healthcare, where outbound communications are used for both preventive care and hospital discharge.

What can you learn from tuning data? Once your application is deployed, what can you learn from callers’ reactions to it? Tuning data from your production system can demonstrate design success and help identify specific interactions that should be improved by analyzing actual callers’ reactions. This session provides an overview of tuning and walks through real-world examples to demonstrate the value of tuning data in demonstrating successes and failures with design, grammars, or speech parameters.

Achieving a truly natural, conversational voice interaction has been the Holy Grail of the speech technology industry for decades. This panel analyzes the progress that’s been made as well as instances where we have not yet met the goal. Panelists discuss the technology behind natural language and customer perceptions of interacting with natural language applications and identify challenges and implications for deploying natural language.

When given a choice in a multimodal application, 30% of users use multiple modalities in a single interaction, so it is now imperative to provide customs with the choice to interact using any modality or channel that is convenient. This interactive session starts by setting up a series of design problems for multimodal applications and then goes on to interactively arrive at possible solutions. Attendees learn that some very obvious solutions are not so obvious after all.

This session focuses on best practices from the AVIxD Design Guidelines Working Group on natural language understanding (NLU) technologies and their role in VUI design. What are NLU technologies? What information is available in the scientific literature to guide VUI design decisions when using NLU? What VUI design practices are associated with successful use of NLU technologies? The session, led by a member of the AVIxD guidelines team, encourages significant audience participation in the discussion.

Join us for a hands-on discussion and demonstration of considerations for making, processing, and even editing recordings from members of the AVIxD Design Guidelines Working Group. Download the audio prompts to work with from http://videsign.wikispaces.com/ST2013DIY, and bring your laptop to participate in hands-on exercises in inflection, co-articulation, pausing, and helping you manage your application prompts. The website also has links to free sound editing software if needed.

What’s new in the voice interaction design profession? How have the proliferation of mobile devices and the emergence of digital personal assistants changed the job of the designer? What skills have become vital, and what knowledge gaps do we face? Join us as this panel of seasoned professionals weighs in on the challenges and opportunities facing voice interaction designers today.

In this session, leaders from the Voice Expressivity and Emotion Group (VEEG) of the Creative Speech Technology (CreST) Network in the U.K. discuss the findings of a recently completed 2-year interdisciplinary study of computer speech. Learn about the psychological, ethical, and philosophical issues to end users around machine emotion, the effects user emotion have on the machine and vice versa, and the practical benefits and limitations of machine-generated emotion.

The promises of speech technology are often discussed, but honest discussions of whether these promises are borne out in the real world are scarce. The purpose of this panel is to remedy this by bringing together organizations that have deployed speech applications. Join us to hear about the benefits of speech, as well as the pain points and surprises that always pop up during deployment.

Most companies claim to care about customer satisfaction but are vague on the details about truly understanding the experience being provided to customers and what’s needed to improve it. Panelists in this session explore a wide range of techniques such as customer satisfaction surveys, reports from call center representatives, call observations, and usability studies to help you avoid pitfalls and maximize the advantages of each method.

The past year has highlighted the utility of speech technology for the healthcare industry. The growing popularity of mobile devices and upcoming changes in healthcare regulation have contributed to increased and novel uses of speech for organizations in the healthcare industry. Join us to hear panelists discuss how their healthcare organizations are using speech in new and exciting ways.

We discuss what makes an effective multimodal assistant and some of the challenges that need to be overcome. Then, attendees use a web-based tool and their PCs to build their own multimodal mobile assistant that they can take with them on their phone. We use Angel’s CX Builder, which is a graphical point-and-click tool for building the interaction, so attendees won’t need to know anything technical to build their assistant. For more information go to http://www.angel.com/company/events/speechtek2013/lexee.php.

The W3C has recently published the Emotion Markup Language, which provides a standard that can be used for annotating text with emotion for synthetic speech. This workshop uses the open- source, platform-independent Mary TTS (text-to-speech) system to introduce the concepts of EmotionML. Participants can use Mary and EmotionML to create and save their own expressive short narratives. No background knowledge to complete the project is required. Bring your PC with earphones. Please find details at http://emotionml.syntheticspeech.de.

Have an app in development or an idea for a cool new one? Let us show you how easy it is to integrate Nuance's world-class speech technologies into it. We'll walk through examples using our iOS and Android libraries and briefly discuss the HTTP service. You will also build a simple working speech application that you can take home and use as an example for building your own speech applications. Whether you're an experienced pro or just getting your feet wet with speech, we'll demystify the power of Dragon and make it work for you. For information about downloading software for this workshop and other useful information, please visit http://bit.ly/ZeiugL.

Bring your laptop with earphones, leave with speech-enabled webpages. We start with four demos you can try yourself: composing email, creating calendar appointments, navigation, and searching. Then we take a look under the hood. In this hands-on workshop, attendees learn how to add scripts to their own webpages to create interactive speech experiences. You should have some familiarity with HTML and JavaScript, but expertise is not required. Come prepared as described here: http://goo.gl/P7cUF.