Technically Speaking
Posted: 11/1998
Technically Speaking
Voice Recognition May Soon Unseat DTMF
By Jennifer Knapp
multifrequency (DTMF), the de facto standard user interface for most enhanced
telecommunications services, is on the verge of being pushed aside by speech-recognition
technology.
"Microsoft has demonstrated in PCs (personal computers) the power that control of
the user interface provides," says William Meisel, president of TMA Associates,
Tarzana, Calif. "Their Windows graphical user interface is what most buyers see as
the PC. In telephony, the voice user interface plays that role."
Speech recognition will make possible the automation of tasks by telephone that
couldn’t be automated before, Meisel adds, making the telephone more like an electronic
personal assistant than merely a way to connect to another person.
Speech recognition’s ascent is not imminent. International Data Corp. (IDC),
Framingham, Mass., says the timeline for widespread acceptance of residential
voice-activated services is two years. However, IDC predicts tenfold growth in the market,
from $10 million in 1997 to more than $100 million by 2002.
While it may not be until the next millennium before speech-enabled services find
widespread acceptance in the residential wireline market, improvements in the technology
by vendors is making deployments in telecom–wireless and wireline–more cost-efficient
and feasible.
Better, Cheaper
The laws of supply and demand have driven the price for speech-recognition technology
to an affordable level. "Costs have dropped rapidly, both for the basic [speech
recognition] technology and for the hardware required to use that technology," Meisel
says. Cost is, therefore, no longer a roadblock to deployment, he adds.
Cost reductions are due, in part, to new platforms that can be integrated into existing
networks easily, in a building block environment, so overhauls to an existing system are
not necessary and services can be introduced one at a time or in bunches.
In its discussions with carriers, Compaq Computer Corp., Houston, discovered the lack
of the right tools was preventing carriers from deploying services using the voice user
interface (VUI). "Carriers were looking for multiapplication platforms and tool sets
to allow them to rapidly deploy new differentiated services," says Chris Ebling,
director of the call-processing platforms business unit for Compaq’s telecom network
solutions group. Carriers have been testing equipment in this arena for a few years, but
"they spent a fair amount of money on a single application that was not scaleable,
wasn’t expandable and didn’t support other applications," Ebling says.
Compaq’s VUI-based platforms are outfitted with a "service- independent building
block-based service creation capability," which means, in other words, it plays well
with others and allows a carrier to build new speech-enabled services on the same
platform.
In addition to becoming more cost efficient, speech recognition is becoming far more
intelligent. "The technology is evolving," says Mary Stanhope, director of
product marketing at Priority Call Management Inc., Wilmington, Mass. "It is getting
to the point where you can use more natural speech. Instead of learning the command,
‘Connect,’ you can say, ‘Yes, I’ll speak to Erika.’"
Whether consumers want, or even need, speech-enabled services is not readily known.
However, Meisel says, General Magic Inc., Sunnyvale, Calif., which launched its Portico
VUI-based personal assistant for the wireless market in July, recently announced its
rollout was so successful, it is advancing its advertising schedule to reach a broader
audience. The service, which costs $19 a month, includes long distance, voice mail,
e-mail, address books, calendars and access to Internet content such as news and stock
quotes.
The Quickening
While consumer demand typically is the driver of market growth, experts say it is the
vendors that are hauling this train. Whether a customer likes a new product "really
has nothing to do with whether a market develops," says Walt Tetschner, president of
Tern Systems, Acton, Mass. "We look at markets and where something is developing
based very heavily on what the vendors are doing."
And over the last 10 months vendors are doing a lot–primarily in wireless applications
for the technology; but wireline applications are not far behind.
The core group of vendors driving the development and adoption of speech-recognition
services include Voice Control Systems Inc. (VCS), Dallas; Applied Language Techno-logies
Inc. (ALTech), San Francisco; Nuance Communica-tions, Menlo Park, Calif.; and Lernout
& Hauspie Speech Products N.V., Burlington, Mass. These companies have been partnering
fast and furiously during 1998 bringing the necessary technology to both carriers and
solution providers. (See Vendor Alliances chart on page 84.)
VCS, for example, partnered in March with Finland-based Tecnomen, a designer and
manufacturer of enhanced services systems for telecom networks and paging systems. The
companies worked together to integrate VCS’ speech software platform with Tecnomen’s
enhanced services system, enabling end users to access and control a variety of
applications by using speech rather than a touch-tone interface.
Three months after this pairing, Tecnomen brought its enhanced services platform,
complete with VCS’ technology, to Tele Danmark, the incumbent provider of voice and data
communications in Denmark. Tele Danmark now offers VUI-based unified messaging to its
subscribers using the platform.
ALTech also has advanced the enhanced services arena with its voice-recognition
technology through a partnership with Voicetek Corp., Chelmsford, Calif., a manufacturer
of software applications and platforms for network-based services. Voicetek integrated
ALTech’s SpeechWorks software into Generations, an enhanced services platform. Max-Reach,
an open, modular suite of advanced intelligent network (IN)-based applications built on
the Generations platform, connects to a service provider’s signaling system 7 (SS7)
protocol stacks, communicating with switches and linking to service control points (SCPs)
for IN signaling. This allows service providers to launch applications such as voice
dialing, personal assistant, single number service and unified messaging.
Functions of these services all operate under a VUI interface, which offers:
- Hands-free dialing using natural language commands that allow for differences in
dialect, accent and grammar; - Verbal management of all communications, including accessing messages, screening
incoming calls and routing calls to predetermined destinations; - Call routing to a predetermined destination; and
- Provision for a common message storage area to access to voice mail, fax mail and e-mail
from a telephone, including a text-to-speech e-mail reader.
These speech-recognition vendors, however, are not the only companies pushing to
advance the technology in telecommunications. Heavy-hitters in the computing industry also
are weighing in. Since the beginning of 1998, IBM Corp., Armonk, N.Y., has partnered with
both ALTech and VCS to integrate speech recognition into its telephony solutions. In
addition, Intel Corp., Santa Clara, Calif., made an investment in ALTech to excel the
power of speech recognition processors–an improvement that Tern Systems’ Tetschner says
is vitally important to moving the technology forward.
The Wireless Equation
Acceptance of voice recognition technology in the mobile wireless arena is already
here. Safety likely can be cited as the main reason why the wireless world is seeing more
speech-recognition activity than wireline.
PTT Telecom, the major telecommunications provider for the Netherlands, launched a
voice dialing system in January for its global system for mobile (GSM) network based on a
modular voice processing system from Glenayre Electronics Inc., Charlotte, N.C. The
motivating factor for launching the service was so "PTT Telecom subscribers could
place calls from their mobile phone without taking their eyes off the road or their hands
off the wheel," according to a company press release.
Early advancement in wireless-based speech-recognition technology has brought many
VUI-based services this year for mobile phone users. Bell Mobility and Pacific Bell, for
example, both partnered with Wildfire Communications Inc., Lexington, Mass., to bring
Wild- fire, a voice-activated personal assistant, to their mobile subscribers in Toronto,
California and Nevada. Wildfire offers personalized answering services, which include
voice mail, call display, call forwarding, call routing and voice-activated dialing
through the VUI.
Other VUI applications for wireless include a dictation service from CellNet Data
Systems Inc., San Carlos, Calif., and Speech Machines, Redwood City, Calif., as well as a
voice-recognition service bureau from Intellivoice Communications Inc., Atlanta, and
Premiere Technologies Inc., Atlanta.
"By establishing this service [bureau], wireless carriers will be able to provide
their customers with voice-activated dialing in the short term and other enhanced services
in the future without having to make a large, up-front capital investment," says
Boland Jones, chairman and CEO of Premiere.
Compaq and Northern Telecom Ltd. (Nortel), Mississauga, Ontario, recently announced
their partnership to focus on bringing a voice-activated dialing system to the mobile
market that surmounts the difficulties of a "noisy" wireless environment, says
Compaq’s Ebling. But Compaq is not ignoring the wireline market, either.
"Next year we will be rolling out an advanced intelligent network peripheral
capability," he says. "We are just leading with the wireless voice service
because that is where our strength in the market is. We will target the wireline side of
the market as well with the introduction of capabilities to support the various [Internet
protocol, or IP] interfaces in the traditional carrier market."
Although there have been VUI-based wireline service launches by traditional telcos, the
service has not been prolific. "Enhanced services take longer to deploy," says
Stuart Patterson, president and CEO at ALTech. "They take longer for vendors to
integrate into their systems, and it takes longer for telcos to say, ‘OK, we’re done
playing with this. I’m going to deploy it now.’"
Long distance carriers must be cautious about their deployments to guarantee quality of
service (QoS). So with actionable platforms just surfacing that meet long distance
carriers’ needs, many are only considering the technology at this point, says Andre Kuyzk,
an analyst for Frost and Sullivan, Mountain View, Calif. Consideration of these new
services likely will include a test period before full-fledged services will reach end
users, he adds.
Image: Speech Tecnology Revenue Growth Worldwide
Frontier Corp., Rochester, N.Y., completed its test period for the technology in July,
launching its SpeedLink voice-activated phone card for businesses. With SpeedLink,
customers dial an 800 number and speak their personal identification number (PIN) to
access VUI-based services such as voice mail, travel reservations, teleconferencing and
standard dialing. The service is not available for residential users, says Randal
Simonetti, vice president of communications for Frontier, because the business market
presently is the "right niche" for this technology.
Webley Systems Inc., Schaumburg, Ill., however, is bringing speech- recognition
technology to subscribers by means of Webley, a voice-activated electronic assistant for
the membership-based Webley Network.
For $14.95 a month, plus line charges for basic services, Webley offers voice-activated
dialing, conference calling, voice mail and call transfers all with spoken commands.
So Long, DTMF
In its report, "The Telephony Voice User Interface," TMA Associates predicts
a hundredfold growth between 1997 and 2003 in revenues for advanced speech technology
products and services in telephony "that would not exist except for speech
recognition," Meisel says. While Meisel admits he is a little wary of this
prediction, he cannot help but believe that the demand for the technology will take off.
"One speech-activated service leads to demands for more features that can be added
with a VUI," he says. "A true paradigm shift means that the growth is limited
only by the imagination of product and service developers."
These predictions, coupled with the large-scale vendor activity over the past 10
months, point strongly toward the introduction of VUI-based services in the residential
wireline market. The fact remains, however, that these are visions that only will meet
reality in years to come. Will there come a time when the telephone key pad will go the
way of the rotary dial? We’ll see.
Jennifer Knapp is news editor for PHONE+ Magazine.
