Not very long ago, in an article on segmenting the speech recognition
market, I revealed my desire to talk to every last one of my major
appliances, and to never have to push a button ever again. Well, I'm still
pushing buttons, and even though we're just months away from 2001, none of
my computers is named HAL. From a business standpoint, however, we're
seeing more and more speech recognition markets emerging, which hopefully
will drive the technology and let me talk to my car ... the real love of
my life.
THE BROAD VIEW
For those of you loyal readers who don't quite remember the markets in
which speech recognition is segmented, here's a refresher. Speech
recognition applications fall into three categories:
1. PC-based applications. These applications free the user
from the unwieldy keyboard, and let him swear at the computer instead of
just mashing the keys (the layman's version of the warm boot). An
excellent example of a PC-based application is Dragon's speech-enabled
word processor.
2. Applications embedded on common appliances. This is where
you can tell your oven to cook your turkey or even ask your car for
directions. I personally don't know what I would say to my oven since we
haven't been on speaking terms for several years, but I look forward to
many enlightening discussions with my car.
3. Telephony-based applications. This is the biggest market,
and far and away the most interesting for those in our industry. It also
needs a bit of segmenting in its own right.
THE CONFLUENCE
The largest speech recognition market, which consists of telephony
applications, may be understood as the confluence of four major
categories. We will review each of them in the sections which follow.
Enterprise Applications
Some examples of these products are auto-attendants that use speech rec.
This is basically a replacement for a receptionist, although the companies
that supply these products say auto-attendants may allow the receptionist
to use time more efficiently, rather than suffer a layoff. Either way,
these products allow the user to ask for particular people or departments.
Other examples include speech rec-enabled voice mail, allowing users to
navigate by means of vocal utterances instead of pushed buttons. I know I
would love to be able to say, "Play my voice mail and then transfer
all of my calls to Mike." (Mike's the guy in the office next to me.)
These are compelling applications, but they may fall short of achieving
"must have" status, at least as far as most organizations are
concerned for the next few years. I am forecasting only moderate growth in
this area.
Call Center Applications
These products are where speech rec has the greatest depth, and for very
good reasons. The biggest expense in running a call center is manpower,
and any application that can cut that cost can save a company millions.
Examples of these products are automated stock quotes, airline
reservations, and other menial activities. It is very likely that speech
rec will change the demographics of the typical call center galley slave.
Simple tasks will be relegated to the speech rec system, and humans will
be used only for more difficult interactions. This market has been big in
the past and will likely continue to grow at an even faster pace.
Service Provider Applications
There are plenty of service providers out there, and their pockets are
very deep, which is why speech rec companies target these markets.
Applications include directory assistance and automated account
maintenance (which might include such tasks as recharging a calling card).
This market is large, but probably won't experience rapid growth in the
near future.
Voice Portals
Voice portals are the wild card in speech recognition. The idea is that
you can call one of these services and gain access to multiple businesses
and information sources in the same manner as through the Internet.
Examples of voice portals can be found at www.bevocal.com
and www.tellme.com.
Voice portals represent a very cool idea, and they've received quite a
bit of media attention. Many of the speech rec vendors are betting their
future that this market will explode. Personally, I think it's a good
idea, but that it won't take off quite as quickly as the media would have
us believe. Many of these sites are still a bit limited. Most importantly,
when 3G wireless arrives, it will be easy to browse the Web at high speeds
using a cell phone. These technologies -- voice portals and 3G -- are in
direct competition.
However, until 3G arrives in mainstream, voice portals do their job
very well. I've gotten into the habit of using a few of them for
restaurant recommendations, stock quotes, and move times while I'm on the
road. If I use my car's speakerphone, then I can pretend I'm talking to my
car ... just the way I always imagined it. Forecasting this market is a
tough bet, but I'm predicting moderate to large growth for the next
several years, and then a slowdown as 3G takes over.
SPEECH REC WATERSHED
Even with all of these great products, here and on the horizon, I'm still
pushing buttons, both on my computer and my telephone. I do believe that
the telephony applications will be the first to achieve any significant
impact among mainstream consumers, and that talking
cars/refrigerators/televisions will be relegated to niche markets for the
techies for some time to come. However, within the next two to three
years, it will become commonplace to use speech rec when calling an
office, call center, or service provider.
Brian Strachman is industry analyst, Voice and Data Communications,
Cahners In-Stat Group. To correspond with the author, please send your
comments to [email protected].
[ Return
To The November 2000 Table Of Contents ] |