Speech recognition technology has come a long way in a short time. The challenge is creating software that can act as an interpreter between the speaker and the computer, converting natural-language inquiries into commands that computers can execute, and translating computer output into voice responses that are meaningful to the caller. To do so, the software must be able to recognize a wide vocabulary spoken by many different speakers, something that the exponential growth in computing power has recently made viable. Nuance Communications’ solution comes in the form of SpeechObjects™, reusable speech components that incorporate the potentially complex dialogue and vocabulary that may be used in speech applications. Due to the modular nature of SpeechObjects™, they don’t require a developer to have specific speech or linguistics experience, reducing the expertise required to develop speech-based applications. Based on open industry standards, such as Java or ActiveX, they can be customized, assembled or created for deployment on major server and Interactive Voice Response (IVR) platforms. Developments such as SpeechObjects™ are significant because they allow organizations to implement a natural speech recognition interface company-wide, without having the complication and expense of contracting developers with speech recognition experience to implement the technology.
<TOP>
VoxML
According to the Simmons Market Research Bureau, 54% of US businesses now have websites to provide information to their customers. That’s great for the people who are comfortable with personal computers and have access to the Internet, but what about everyone else? Motorola’s answer is VoxML Voice Markup Language, a programming language that allows access to online content from the Internet or intranets using any telephone. "This language enables developers to leverage their existing web application skills to tap into new distribution channels and reach emerging markets and revenue streams." VoxML is a language for interfacing voice requests with the basic language of the Internet, HTML (Hypertext Markup Language), which is text-based. Applications based in VoxML can take a voice request and translate it into a web request. It can then locate the content requested and publish it in VoxML, which can then be translated into speech and delivered back to the user. VoxML is capable of interfacing with encapsulated dialogue components such as Nuance’s SpeechObjects™, allowing developers to create a natural-speech interface for callers making inquiries and tap into existing web content to provide the answers. "It leverages the work done gathering information at websites by making it much easier to access that information by phone," explains Bill Meisel, a consultant with over 20 years experience in speech recognition and the president of TMA Associates. "Wireless and landline telephones can become a convenient way of getting information from the Internet, no matter where you are." Several content providers have embraced VoxML to extend the reach of their online services, including CBS MarketWatch.com, Traffic Station Group, Biztravel.com, astrology.net, SmartRoute and The Weather Channel. V-Commerce Bringing together these and other innovations allows consumers to serve themselves electronically through multiple channels, while allowing businesses to leverage their existing resources to a wider audience and centralizing information gathered on clients for a greater ‘one-to-one’ approach.
<TOP>
Basically, there are three v-commerce strategies:
Speech In , Speech Out - with v-commerce, the telephone or cellular phone offers consumer-friendly access to a range of commerce applications at an extremely low cost. The applications typically involve a dialogue over the phone with a speech system, such as SpeechObjects™. This solution is well suited to transactions that would typically take place today using a touch-tone keypad, or by speaking to an agent in a call center.
Common Business Server - As businesses move to offer more services over the web, they are also realizing the need to offer the same level of self-service over the phone. This is achieved through integrated business and speech servers that support transaction execution through both graphical and voice interfaces. This not only extends an organization’s investment in web content to phone users, it also enables companies to warehouse the information they gather in interacting with customers in one location, no matter the channel through which that contact took place.
Speech In, Text Out - The integrated use of voice input and visual output raises the bar for efficiency in completing a transaction. Instead of clicking through multiple web pages, users simply speak their requests, and instead of the response being read back to them, it is displayed for easy reading. Even for existing personal computer users, this could make self-service on the web much more appealing. Of course, an organization might pursue all three strategies, or focus on one, depending on the nature of their business.
An example of a typical application concerns travel arrangements. A person might log on to an airline’s web page and order tickets for a flight to New York. On the day they are supposed to depart, a storm hits the city, and they want to know the status of their flight. They call the airline’s toll-free number and ask, "Is the flight to New York leaving on time this afternoon?" The v-commerce system, realizing it needs further information, might ask the caller’s name or a code given during the reservation of the tickets. After the caller has provided the required information, the system searches the airline’s web page and notes that, currently, the flight is delayed and travelers are cautioned to call in for further updates as the storm progresses. This information is converted back to speech and relayed to the caller, who can either ask for more information or terminate the call. Meanwhile, the system records that the customer has called for this information in the same place that the original information about the reservation is recorded, allowing the system to ‘remember’ what has transpired and individualize future exchanges.
<TOP>
Advantages
This approach has a number of advantages for the consumer over traditional e-commerce. First of all, it makes self directed electronic transactions more accessible, allowing customers to use the device that is most convenient at any given time. V-commerce is easy to use - callers simply dialogue with the system as they would a live attendant - so there’s no new skills to learn, no special knowledge required. It’s affordable - there’s no need for users to buy special equipment because any phone can access a v-commerce system. V-commerce also offers a security advantage over text-based interfaces. Voice interfaces can use a customer’s unique voice print to restrict access to secure information, making transactions more secure and helping overcome a concern for traditional e-commerce users. For businesses, v-commerce offers the opportunity to leverage their existing investment in web content to provide self-service electronic service to more people through multi-channels. It also rationalizes the interactions through centralized servers, making it easier and more practicable to build relationships with customers by individualizing interactions. While this does require some additional investment in v-commerce solutions, most are designed to be simple to implement without a large amount of specific expertise. As a result, customers can more frequently serve themselves and the organization manages to provide an improvement in service with a reduction in cost. As technology improves, it is possible to do more with less knowledge. The first steps with e-commerce focused on an audience of computer-literate web surfers. Now, with the advent of voice-based systems, electronic commerce can be available to anyone who has access to a phone. Not only will this take e-commerce to a wider audience at a lower cost, it will also give consumers far greater access to the companies that serve them - from their location, on their schedule and through the most accessible and preferred channel.
<TOP>
|
