September  2000

BY MUKESH SUNDARAM, TELERA

Outsourcing of voice telecommunications is a mature and prevalent practice in the industry. With the expansion of Internet technologies into the world of telecommunications, new outsourcing options are being made available to businesses that allow them to retain control of applications while still outsourcing much of the headaches and expense. We will explore how the Internet is changing telecommunications outsourcing and discuss what advantages this trend can bring to your business.

Businesses tend to outsource telecom to service providers that can undertake the entire application on their voice hardware infrastructure or that can add value to the path of the call such as routing the call appropriately between different business locations. The dominant players in telecom outsourcing are long-distance carriers like AT&T, WorldCom and Sprint. Other outsourcing agencies, such as Call Interactive and Price Interactive, focus on managed voice processing infrastructure to host self-service applications for businesses.

When businesses outsource voice response applications, the customer-specific data that are used or generated during the voice interaction are often batch-transferred to the outsourcing entity. The functionality is typically constrained because businesses transfer only the least amount of data required by the application that has been outsourced. Customer data are the most precious assets of businesses and are fiercely protected; hence, the types of outsourced applications possible are constrained by the unwillingness of most businesses to entrust extensive customer data to an outsourcer. Furthermore, since the platforms upon which such applications are built use proprietary development environments and languages, changes are usually slow and daunting due to the level of interaction necessary with the outsourcer to achieve them.

A more recent development in the speech technology industry is voice recognition. This technology has matured and is now more easily deployed. It is, however, still quite expensive and requires fairly heavy-duty systems to perform the recognition. This is another reason businesses outsource voice applications -- they do not want to incur the cost of the expensive systems, voice processing hardware and recognition technologies.

New Imperatives To Outsource
With the Internet becoming the vehicle for customer contact and innovation, business is changing from being inwardly focused in its IT infrastructure to becoming more customer- and partner-centric. Customers are more demanding and expect more personalized treatment. Loyalty and retention of customers is achieved through excellent and differentiated customer service, which, to a large extent, hinges on personalization of service.

Customers want to reach businesses through a variety of channels. The traditional voice channel is being augmented by new access methods using PCS and wireless devices. Despite the addition of new channels, the customer almost always resorts to voice when it comes to customer service. The standard telephone is still the most ubiquitous device available for contact.

Businesses must create multichannel infrastructures. With conventional technologies, the voice hardware silos connect to the back-office data through their own means. Web infrastructure uses a different technology path to get at the same data. Over time, this divergence leads to maintenance problems. A fix to one path does not benefit the other. Web platforms also have the undivided attention of the tools, data mining and personalization software industry, to name a few. Clearly, it is beneficial to the business to create a single infrastructure to support all interaction channels.

The term "Internet-time" has been adopted by businesses to mean agility in making changes to their business processes and the way they communicate with customers. Ease of modifying Web applications and having any logic change take effect for the very next contact is the level of flexibility and control desired by businesses in this new climate.

Web Technologies Drive New Methods Of Outsourcing
In telecom, this separation has occurred in proprietary ways in certain vendor products. Usually, the voice hardware host is separated from the computer that executes the application determining the logic of interaction with the caller. More recently, XML (extensible markup language) has become a means of separating the voice infrastructure from the application, just as HTML achieved this separation between the browser and the Web server. The W3C, which is an Internet standards body, has now accepted a candidate XML specification for consideration as the Internet standard in its Voice Browser Working Group. This specification is called VoiceXML and is the product of the VoiceXML Forum.

Widespread adoption of this standard means that voice telephony equipment manufacturers can separate their interaction platform from the Web application server platform. This will fundamentally change the way voice applications are created over the next several years. Businesses will have more product choices when deciding to build and manage their own infrastructures. This will be a marked change from today's practice, where a customer must select the same vendor for both the hardware and software for voice response systems. Voice response system vendors today lock their customers onto their proprietary platforms, ensuring a dependency by the customer on the vendor to add new speech technologies to the platform.

With the maturing of Internet infrastructure, availability of bandwidth and the use of XML-based separation of application and voice hardware, it is now possible to separate across a wide area the voice telecom hardware from the application. This enables a whole new way of outsourcing the voice hardware, while retaining control over the actual application, which is merely a Web server emitting XML pages. The outsourcer hosts and manages the voice hardware in any location of its choosing, and callers interact with this hardware. The voice port communicates with the Web server deployed by the business, which instructs it to interact with the caller in a certain manner.

Moving the voice hardware out of the business premises, while retaining control over the application logic, is the new paradigm in voice telecom outsourcing. This approach reaps the benefits of outsourcing while maintaining control over the caller experience. Maintaining control means that the business can change the caller experience and have it take effect on the very next call. Furthermore, since the application logic executes on a Web server at the business premises, this allows for the reuse of back-office integration software modules which are heavily customized to the business. These back-office integration modules are often the most challenging part of any Web or voice application functionality, and a unified approach for both channels saves the business significant money.

This new means of outsourcing telecom makes it easy for the business to protect customer and other sensitive databases from the outsourcer's environment while providing just the data needed for a particular caller. If necessary, the communication between the Web server and the voice hardware can ride on top of SSL, which is considered to be sufficient security for the Internet.

Challenges For Businesses Transitioning To Web-enabled Telecom
A number of new players are now on the landscape, promising to take away the pain of maintaining a premises-based voice infrastructure. These new players have fundamentally adopted XML as the platform control language. Others who have been in the business longer have also begun to migrate to XML using their existing commercial platforms.

The challenge businesses now face is how to select between these new alternative outsourcers. When outsourcers are being evaluated, it is important to understand their core business focus. Equally important to understand is the reliability and redundancy of the architecture underlying the hosted telecom infrastructure. The business should also assess whether the application is mission-critical, and if so, how the outsourcer supports this. Another consideration is how the infrastructure handles the security and privacy of the data the business sends to the outsourced platform. As with any Internet-based architecture, there is exposure to the same types of attacks that many public Web sites have suffered recently. If the outsourced infrastructure is mission-critical, this is a crucial consideration.

At this time, we are at the dawn of a new era of Web-driven telecom outsourcing. Advances in the use of XML to command and control voice infrastructure have been proven over a wide area to deliver mission-critical, business-controlled, caller-personalized applications. Standardization of VoiceXML by the W3C will expand the choices available to businesses for both on-premises and outsourced voice infrastructure.

A unified Web development platform for customer-facing applications enables businesses to reduce their total development costs, while leveraging personalization technologies long available to browser-oriented Web applications to deliver an enhanced voice experience to their customers.

Mukesh Sundaram is the guiding technical hand in development of Telera's patented Intelligent Communications Network architecture. Telera, a privately held company located in Campbell, CA, enables businesses to deliver personalized telephone-based customer applications designed to extend existing e-commerce and customer contact solutions to the very edge of the network. Telera offers value-added services to enterprise customers on a local access, wide-area, distributed, private IP applications network.

[ Return To The September 2000 Table Of Contents ]