By Brad Smith
December 4, 2000
Better Voice Recognition Hits Market, Sparking Consolidation
If press releases were chad, election officials in Florida
would have had an easy time toting up the winner in the wireless
race this fall. Except the victor in this race would have
been voice-access technologies.
Carriers like Qwest Communications International Inc. and
Sprint PCS and Internet portals like Yahoo! Inc. have made
announcements about end-user services. But most activity took
place in the rollout of new products and services from a host
of vendors. And a wave of consolidation has begun to take
shape that is certain to overtake the marketplace.
The mounting interest in voice access is due in large part
to the increasing sophistication of that technology. And that
has led carriers and others to realize they can make considerable
amounts of money by offering voice access to a variety of
services. Allied Business Intelligence, in a recent study
on voice recognition, forecasts that so-called vCommerce will
be a $50 billion market by 2005. The consultancy says that
in North America alone mobile voice portal users will increase
from about 1 million in 2001 to more than 56 million by the
end of 2005.
Many of the companies offering voice access have focused
on the markup language standard known as voice extensible
markup language. The first version of the open standard, under
the aegis of the VoiceXML Forum, was published last May. Like
all emerging standards, it isn't perfect and now companies
are looking to surpass what it offers.
VoiceXML is the outgrowth of an effort put together by AT&T
Corp., IBM Corp., Lucent and Motorola. Those companies wanted
to foster a single standard to make Internet content accessible
by voice and phone.
Much like criticisms levied against the Wireless Application
Protocol on the data side, some view VoiceXML as limiting
because it requires that voice sites be written in a language
separate from the Internet (hypertext markup language) standard.
Why not give voice access to HTML directly?
Among the companies looking to do that are Talk2 Technology
of Salt Lake City and VocalPoint Technologies of San Francisco.
Two-year-old Talk2 recently completed a $43 million venture
capital investment round and just now is emerging from a stealth
mode with its core technology, which it calls ViPrNet. The
company sees carriers, both wireless and wireline, as its
main customers and currently has trials with some carriers,
according to Kary Burns, market development vice president.
VocalPoint acts as both an application service provider for
voice access as well as licensing its VoiceBrowser technology
to carriers and portals. As an ASP, its first customer was
a health-care benefits administration company, @bovehealth,
and it recently started working with Telecom Italia to voice-enable
that carrier's "ProntoWeb" portal.
As a core technology, Talk2 wants its ViPrNet architecture
to work with any enabling protocol or solution, Burns says.
It can use voice recognition from many vendors, and also can
work with both VoiceXML and HTML Web sites as well as WAP.
The company believes that voice-enabling normal HTML sites
makes sense as long as VoiceXML is in its infancy.
Burns says Talk2's competitive advantage is that its technology
is integrated into the network backbone, giving it scalability,
flexibility and robustness for both carriers and enterprises.
He says the technology can voice-enable any existing information,
including e-mail, calendars and contacts, without having to
modify the source but also providing the tools to create VoiceXML
sites.
VocalPoint has a different focus‚voice-enabling existing
Web content either by licensing its VoiceBrowser or acting
as an ASP. Its browser works with both HTML and extensible
markup language.
Given the intuitive nature of voice access, these developments
are certain to garner a lot of votes.
|