A topic of great interest and controversy in the speech industry is competitive benchmarking, studies designed to determine which of two or more recognizers is most accurate. Unfortunately, it is easy to produce skewed results, either by mistake or deliberately, so that the result are misleading. We will discuss common errors in experimental design or analysis and ways to ensure that tests accurately correspond user experience. "

Analytics tools for IVRs and contact centers have improved significantly during the past decade. However, when we use these tools separately for self-service systems and contact centers, we miss the opportunity to optimize overall caller experience and improve automation rates. This session discusses various methodologies and tools for cross-channel analytics that can help identify caller intent. Attend this discussion to learn how analytics can improve ROI as well as customer satisfaction metrics and the overall customer experience.

Speech and language disorders affect millions of adults and children and can greatly affect an individual’s quality of life. Speech and language technologies have a huge potential for addressing speech and language disorders, both for rehabilitation as well as for assistive technology. Software incorporating linguistic technologies can provide objective, consistent, and tireless feedback to language-impaired users. We discuss existing software, the principles of using speech and language software for language deficits, how different types of feedback apply to specific disorders, and what new applications might be possible.

What are the pros and cons of using freeware platforms such as Asterisk to deploy speech applications? What challenges and hidden expenses are associated with deploying and maintaining a speech-enabled IVR solution? Other issues include using a freeware toolset to develop, test, deploy, and monitor an application; scalability issues; poor documentation; and the availability of competent resources to create a world-class solution. Learn the total cost of using a freeware speech platform compared to commercial platforms.

Prepare to laugh and learn as a decade of Wynn Solutions research reveals what makes owners and managers of top-performing businesses so effective—and it’s not what you’d expect! If you enter this keynote session expecting to learn all about best practices and superior strategy, you’d better read the title again. Standout success in a challenging economy often comes from personal advantage, whether it’s perceived value, personality, leadership style, communication skills, abundant resources, or access to privileged information. This entertaining session takes a realistic look at the influence you’ll need to make sure your knowledge and hard work deliver the results you want.

Join us for a discussion of the benefits of creating powerful, next-generation user interfaces that combine natural language processing, omnichannel prediction, Big Data, and real-time machine learning. Learn about a predictive natural language (PNL) solution that leverages a powerful statistical model to integrate interaction data across multiple modalities to accurately predict and understand user behavior and intent. By contextualizing generic language models, PNL solutions outperform traditional natural language interfaces and improve task completion and self-service rates.

Human-assisted understanding is a new hybrid approach to speech recognition in which humans take over some recognition tasks from the system. Systems that rely on human-assisted understanding hide recognition errors from the user, resulting in speech interactions that appear to be more natural and robust, but are there drawbacks to this approach? This session explores the benefits and limitations of human-assisted understanding in creating a positive and productive customer experience.

The core technology underlying all current speech analytics products is either phonetic-based or word-based. This presentation offers a rigorous yet accessible explanation of how these two approaches work and provides an objective assessment of their benefits and shortcomings, as well as a detailed comparison of both technologies along the key dimensions of accuracy, speed, and storage requirements. This session arms you with the knowledge to make more informed decisions when faced with these two competing technologies.

We are in the midst of rapid change in the speech industry, but is this a true paradigm shift or an adjustment to the status quo?Join us as panelists explore how changes to companies in the speech market, as well as new opportunities in wearables, home automation and entertainment, automotive, and mobile will affect future opportunities in speech technologies.

Many organizations collect customer feedback about their speech applications and other elements of the customer experience, but often lack guidance on making the best use of this data. In this session, you hear three success stories in which the appropriate use of customer feedback led to better designs and better decisions about the speech application. These lessons will help your organization optimize the value of customer feedback to yield better customer experiences.

User authentication is a necessary first step for a vast variety of tasks. In spite of the ubiquity of authentication, it remains a challenge for automated systems to verify who is on the line before allowing a user to access personal data or perform transactions. Panelists in this session share their experiences with difficult authentication situations as well as tips and techniques for overcoming them.

Analytics has received a great deal of attention in recent years, with just cause. Analytics provide new tools and methods for manipulating large amounts of data from multiple sources, which may result in dramatic cost savings, solid metrics, and eye-opening results. Get prepared! In this presentation, learn the basics of dealing with analytic data, methods for avoiding common stumbling blocks, and hear lessons learned from real deployments to help you ease into your first analytics effort.

Each call to a speech IVR is a conversation that, like human conversations, can easily be disrupted when one party misreads the inherent turn-taking cues and talks out of line. While humans easily recover after interrupting each other, a single turn-taking error can derail a conversation with a speech application. This session explores the concept of turn-taking and offers actionable advice on how to prevent, diagnose, and repair designs plagued by turn-taking issues.

The handoff of speech projects from the design team to developers is often less than successful because of poor communication and documentation. This session remedies this situation by providing a set of “spec-writing” guidelines to make developers’ jobs significantly easier by allowing them to focus on coding the application instead of deciphering the design specification. Such improved communications help to avoid project setbacks, reduce rework, and save time and money.

Personalization allows us to offer customers a more relevant interaction by utilizing the vast amount of information organizations have about them. Designers often limit the use of personal data for fear of triggering users’ concerns about Big Brother, but are we unnecessarily restricting design creativity? Attend this panel for an exploration of the circumstances when customers are creeped out by technology and what it knows about them and finding the line between creepy and cool.

The solution to reducing driver distraction likely resides in combining a variety of driver interfaces to fit specific tasks. Speech plays a critical role for text entry while driving, but holistically, complex infotainment systems require simple, multimodal user interfaces. This presentation describes a multimodal HMI design model, along with supporting research, for driver actions such as task selection, list management, text entry, making requests, and map control, which are required to perform a growing assortment of in-vehicle, nondriving tasks.

Multichannel user experience has been a hot topic for a number of years, but integrating voice user interfaces in multichannel and cross-channel customer journeys remains a persistent challenge. Join panelists in this session as they analyze where voice brings the most value in multichannel interactions, discuss best practices for creating optimal voice user experiences, and share their experiences to help you design great multi- and cross-channel user experiences.

This panel explores the opportunities and challenges associated with the digital healthcare revolution and the role of speech technologies in this changing world. What are the biggest obstacles to digital healthcare becoming a reality? Where do speech technologies bring the most value to healthcare? How will health providers, insurers, and payers provide patient support in the world of digital healthcare? Join us for analysis of the state of digital healthcare today and predictions for its future.

This session tells the story of a major healthcare organization as it decided to build its initial SMS service. Hear about its immersion into the unexpectedly complicated world of SMS messaging from short codes to the ultimate question of providing a service versus a solution for customers. This presentation gives common-sense advice on how to approach SMS as an integrated service channel that builds both positive customer experience and solid business value.

Speech technologies are playing an increasingly important role in healthcare delivery, monitoring, and communications. Attend this unique session to hear a series of 6 minute lightning talks on how speech is playing a part in transforming today's complex, inefficient healthcare into a better system for both patients and providers.

Experts describe recent advances and results that will soon be available for building products and services. Product planners and technology buyers can benefit from attending this session. Topics include speech recognition (new paradigms such as the template approach and deep neural networks), natural language understanding, dialogue systems (structural and machine learning approaches), and new applications (speech summarization, meeting recognition, media stream search, and others).

Omni-channel is the new multi-channel: an interconnected experience where customers can pause and resume transactions as they please, jump seamlessly between channels as well as between self-service and agent-assisted service, while context and history is always preserved. The new mantra of the mobile customer experience: I shall never have to repeat myself again. This session will discuss the challenges of going omni-channel in customer care and what architecture needs to be in place to deliver on the vision. It will describe an architecture that allows for optimal reuse of resources and IP already created for IVR and integration between the different channels while leveraging each channel’s unique characteristics.

Artificial Intelligence Markup Language (AIML) for Virtual Assistants 2.0 is a popular, standard technology used to create talking artificial intelligence characters (chatbots or virtual assistants). AIML 2.0 contains learning features so that the app can be taught personal information about each user. The language specification also includes a collection of functions to implement device control actions such as dialing, texting, sending and receiving emails, opening apps, checking device status, and more. Attendees use AIML 2.0 to build a simple virtual assistant and learn how to voice-enable it. See http://callmom.pandorabots.com/static/speechtek for details.

In this continuation of D103, learn how to add speech recognition and speech synthesis to your personal assistant using the Android software development kit. Your virtual assistant will be able to chat about a range of everyday topics. You learn how to add the ability to execute search queries, control device functions, and make calls to web services. Take home your own speech-enabled virtual assistant. See http://www.spokenlanguagetechnology.com/speechtek/workshop.html for details.

We have tuned the speech recognition accuracy of many IVR applications and significantly improved the performance by simply tuning speech grammars and confidence levels, with little or no changes to the application call flow. In most cases, we have been able to cut down the speech recognition error rate by around 50%. How do we do it? Come and learn the key tuning techniques that we have developed through the years which have enabled us to achieve such consistently good results.

See two 20-minute presentations about some of the latest technological developments in speech technology from companies at the conference.

Join your peers on Monday after the sessions for the opening of the Customer Solutions Expo - featuring the CRM Evolution, SpeechTEK, and Customer Service Experience Showcases. Visit with conference sponsors, exhibitors, speakers, and other attendees while enjoying light hors d'oeuvres and drinks.