The
Voice Web: Commercial Implications of VoiceXML
The
Voice Web is convenient shorthand for the extension
of the graphical Web to voice, particularly to wireless
and wired telephones. The Voice Web is made practical
by speech recognition, supported by text-to-speech synthesis
and speaker verification in some applications. Speech
recognition allows a flexible and more natural user
interface, removing the limitation of the touch-tone
keypad as the primary telephony user interface.
We
are accustomed to viewing the Web via a Graphical User
Interface, typically a Web browser on a PC. It is thus
easy to conceptualize the Web as being what we see.
The Web consists of a distributed data, most of which
is interpreted through a graphical Web browser. While
much of this data is currently represented in HTML code,
in the future the data will be separated from presentation
style using XML technologies. One can also think of
scripts that describe rules for handling the data and
user input as another form of data.
The
Voice Web is a way of using that data and those rules
through a voice dialog. In most cases, the way we will
get at that data for voice will be quite different than
how we do so for a graphical Web browser, but the data
is the same. With this broad view, the commercial importance
of the Voice Web will become more obvious. As we will
see in the remainder of this article, the Voice Web
is much more than just an extension of the visual Web.
VoiceXML
and the Voice Web
VoiceXML is as important to the Voice Web as HTML was
to the development of the visual Web. It provides a
"voice view" of Web data and transactions
that is independent of the various "voice browser"
implementations that interpret the VoiceXML and the
speech engines and telephony interfaces that underlay
the voice browser. The costs of running the voice browser
can be supported by a "voice portal" that
is separate from the site on which the VoiceXML code
resides-just as a user's graphical Web browser supports
the cost of rendering HTML on the user's computer screen.
Another
key advantage of VoiceXML is its simplicity. A large
number of developers can readily master it, particularly
those that have web development skills. Its support
by many speech technology vendors will also encourage
development of higher-level development programs that
compile into VoiceXML, but allow the specification of
dialogs in more friendly ways (as has occurred with
HTML).
The
Voice Web is growing rapidly, even without the full
maturation of the VoiceXML language and VoiceXML interpreters.
Nevertheless, a standard is critical for the Voice Web
to meet its full commercial potential. At the moment
the VoiceXML language specification is progressing under
the efforts of the W3C Voice Browser Working Group.
Business
Opportunities
There
is a hierarchy of opportunity created by the Voice Web
(see Figure 1). A company can occupy one or more of
the segments shown. For example, a voice hosting company
can simply provide a platform for delivering voice services
that is shared by many companies for efficiency. That
same voice hosting company is also likely to offer professional
services to help customers develop applications. It
may even develop proprietary tools that accelerate development
to give it an edge over rivals.
Content
|
Information
& E-Commerce Services
|
Professional
Services
|
Tools
|
Hosting
|
Application
Software
|
Middleware
|
Transport
|
Platform
|
Speech
Technology
|
Figure
1: Segmenting Commercial Voice Web Activities
VoiceXML
probably fits best in the middleware category; in the
sense that it provides a platform that isolates application
software from the speech technology and voice transport
(the telephone network). Nevertheless, it impacts all
the other categories. Vendors in these categories must
decide if they want to support VoiceXML, or perhaps
even build their business based on VoiceXML. For example,
a vendor of application software must decide whether
to build applications in VoiceXML or use another, perhaps
proprietary, application programming interface and a
standard programming language.
To
those cognizant of the long-term advantages of VoiceXML,
the choice might seem obvious. But, today, VoiceXML
developers tend to rely on proprietary extensions to
take full advantage of features in most technology vendors'
software. This will change as the technology stabilizes
and VoiceXML matures, but when to adopt it is currently
a difficult decision for many developers
Continued...
back
to the top
Copyright
© 2001 VoiceXML Forum. All rights reserved.
The VoiceXML Forum is a program of the
IEEE
Industry Standards and Technology Organization
(IEEE-ISTO).
|