VOICEXML - VOICING THE WEB

Introduction

The customer is the cynosure of 'New Economy'. In this era, customer service is 24x7x365 and customer service infrastructure has become a key part of many businesses around the world. Customers expect service and information about services, anywhere, anyhow, anytime and on just about any communication device they carry. Consequently, enterprises strive to keep the elusive lot by exploring and investing in newer and better means of communication and conducting businesses. And most of the action can be seen on the ever-growing Internet, which has offered businesses a new global communications medium for efficient sharing of electronic information and transactions. For example, VoIP (Voice over Internet Protocol) has completely changed the face of services offered over long-distance telephones.

Even after the stupendous growth of the Internet, the telephone remains a viable and popular media for communication and conducting business. Telephone networks are widely and readily available, simple to operate and use the most natural form of communication - the human voice. The proliferation of the cellular phone has made access to the telephone network more easy. Corporations have been using telephone-based technologies such as IVR for communicating with customers and doing business through call centres, telemarketing and help lines. But these are proprietary in nature and made for a particular application using high-level programming languages, leaving very little scope for customisation. Additionally, IVRs use Dual Tone Multi-Frequency (DTMF) for collecting inputs from users that greatly limit the applications they can run.

The need for better access to online business

The Internet as a communication medium is growing everyday with increased availability of personal computers, networks and common standards of software and applications. However, access to the Internet is still restricted, as users must have a computer and a live Internet connection. And accessing the Internet on a personal computer also implies limited or nil mobility. Recently, the Wi-Fi technology has solved many a mobility concerns of personal computer users (accessing the Internet) but still, a Wi-Fi hotspot is required in the vicinity to make a connection, bringing us back to square one!

Mobile access to the Internet is possible through cellular phones. People have been quick to spot this feature, and apparently, the rate of proliferation of mobile devices (cellular phones, Smart Phones, PDAs, etc.) offering Internet access has surpassed that of personal computers. About one in ten people in the world now own a cellular phone! And the number of cellular subscribers is expected to surpass the number of fixed telephones by the end of this decade. Such a transition is already happening in certain countries where penetration rate of mobile phones is around 75 to 80 percent.

Nonetheless, the Internet experience on these devices is hardly satisfying with small displays and clumsy keypad. Entering text, using the standard 12 keys on a cellular phone isn't very convenient for youngsters, let alone the older populace who have to peer through their spectacles to read text on cellular phones. Therefore, the goal of anytime, anywhere, access to an array of information and services still remains unfulfilled.

Enter VoiceXML

Voice eXtensible Markup Language, or VoiceXML, as it is popularly known, is an XML based markup language that enables users to access services offered over the Internet through a telephone! And it sure opens a whole new world of possibilities to people who are not Internet-savvy. Given its open architecture and simplicity, telecom operators, ISPs, call centres and businesses alike can use VoiceXML to provide 'feature-rich', value added services. These services can be delivered through a normal website or database on a conventional wireline or wireless phone without any special protocols such as the WAP or touch-tone alternatives. The markup language strives to bring the full power and advantages of web-based development and content delivery to interactive voice response applications.

How does VoiceXML work?

VoiceXML is essentially the HTML for telephone-based speech applications. Proposed by the VoiceXML forum (http://www.voicexml.org), an industry organisation founded by AT&T, Lucent, IBM and Motorola and currently in its version 2.0, the markup language uses Speech Recognition and DTMF for user input, and pre-recorded audio and text-to-speech for output. VoiceXML can be thought of as a container for dialogs. There are two types of dialogs: forms and menus. While forms present information and gather input, menus offer choices of what to do next. VoiceXML scripts are just like HTML scripts. However, as opposed to HTML, which is executed on a web browser, VoiceXML scripts are executed on a part of a VoiceXML system called the VoiceXML Interpreter (aka. the gateway), which is accessed by a user through a wireline or wireless phone.

The VoiceXML architecture

A typical VoiceXML based system contains four main components:

1. Telephone network: The normal PSTN network, or a VoIP packet network.
2. VoiceXML Gateway: This comprises a VoiceXML interpreter that works with media resources (speech recognition, text-to-speech, audio playback) and telephony resources (DTMF, call control). VoiceXML Gateway downloads the applications from Application Server and interprets it.
3. Application Server: Usually, a web server on which VoiceXML applications and business logic are hosted.
4. TCP/IP network: LAN, WAN or public Internet. VoiceXML connects to the telephone network on one side and TCP/IP network and Application Server on the other side.

The VoiceXML architecture is much similar to traditional web-based technologies and the modern wireless technologies such as WAP. A user can call a voice-enabled web service number (the VoiceXML server) from a phone; giving input to the system. On the server, the media resources component translates the user's voice input, retrieves VoiceXML pages from the Application Server via HTTP and performs actions based on the interpreted VoiceXML page. The gateway may also talk back to the user from the VoiceXML page using text-to-speech and recorded voice.

Applications and opportunities

Some of the important applications of VoiceXML include:

• Next generation calling services such as email-by-phone, email reader, personal assistant service, contact diary, reminder of appointments and meetings.
• Voice activated dialling and voicemail that are already a part of cellular services.
• Content services such as news, horoscope, weather reports, stock quotes, sports scores, movie listings, train and flight status
• Business productivity applications such as conferencing, sales force automation, contact centres and unified messaging
• E-commerce and M-commerce applications such as financial and banking transactions, fund transfer, billing, auctions, travel reservations etc.
• Miscellaneous applications such as change of address information, report lost credit cards, track packages and shipments, gaming and lotteries, etc.

As VoiceXML separates business logic from Voice User Interface, it opens a whole new world of business opportunities called Voice Service Provider (VSP). Users can either develop and deploy voice applications on their websites or outsource them to be developed and hosted by a VSP. These hosted services can be accessed by users through a phone number allotted by the VSP. With growing popularity of VoiceXML technology, several small and mid-size organisations will consider outsourcing their voice-based applications.

VoiceXML is here to stay, as the World Wide Web Consortium has fully recommended VoiceXML 2.0 in March 2004. The recommendation is equivalent to a web standard, implying that the specification is stable, viable for web interoperability, and that the W3C fosters its adoption by the industry. More than 370 companies in the VoiceXML Forum support the recommendation and have developed and deployed VoiceXML-based applications, products and services.

Introduction  |  Contents   |  Top