Speech Technology is Ready For Prime Time!
Speech Technology is Ready For Prime Time!
By Donna Fluss
Congratulations speech market youve finally made it to prime time. It took 20-plus years, billions of dollars in investments and many false starts, but there is no longer a question about the viability or need for speech-enabled solutions.
But you dont have to take my word for it, or the assurance of any other analysts, as many of us have been saying for years that this market is primed to take off. Predictions of great growth are still out there Dataquest recently pronounced that the global voice business market will grow from $656 million in 2002 to $3.1 billion in 2008. And Gartner came in with a more conservative estimate, that the speech software license market is expected to grow from $127.5 million in 2002 to more than $258 million in 2007. Weve seen similar claims for years, but the current projections have a chance of being directionally accurate because of some significant changes in the speech and voice market place.
Heres whats happening:
- User organizations are starting to take interest 2253 people attended the September 2003 SpeechTek show in New York. 66% more people paid for the show in 2003 than in the prior year and 65% of the attendees were end users. In previous shows, the majority of attendees were vendors.
- Microsoft, which has been dabbling in this market for years, has indicated it is going to go to general availability with its speech products. Microsoft has the bankroll, market clout and ecosystem to reach customers who couldnt be reached in the past. If Microsoft puts its marketing machine behind speech, it could invest more in one year than the rest of the companies in the field combined. This can only be positive for the speech market place in general.
- Infrastructure vendors like IBM, HP, Oracle and Intel are taking the speech market more seriously than at any time in the past. With the exception of Intel, which has been serious about speech for years, these four vendors, with combined revenues of more than $206 billion in 2002, are paying much closer attention to the market, probably because of Microsofts entrance. IBM has more than 400 people involved in speech and has been doing R&D in this area for 35 years something very few of us know about, as theyve kept it a secret in the past. HP recently purchased a VoiceXML company and is concentrating on the carrier side of the market. Oracle sees speech as only one of many channels, but one for which it wants to deliver products. And Intel is continuing to reduce the cost of speech hardware, recently announcing a new software solution to eliminate the need for speech cards. If these companies loosen their marketing budgets and make serious attempts to sell their offerings, there will be a significant impact on corporate buying decisions.
- Standards that make it easier to develop speech applications are going to make it easier to deliver them. Now this also means that we are likely to see some really bad speech applications but accelerating the pace of delivery can only help to speed up the general publics acceptance of speech. Ideally, there should be one open programming standard, but because the Microsoft .NET world wanted its own, we now have two competing standards, VoiceXML and speech application language tags (SALT). The competition may, in fact, spur improvements. At the moment, there are very few developers using SALT, but its possible that the threat of this new standard will motivate the painfully slow-moving WCS to quicken its pace in developing VoiceXML.
- Embedded technologies are finding their way into our everyday lives. Speech technology is in our phones and cars and when used well, is so effective that many of us cant understand why its not in our kitchens, laundry rooms and TV sets. Wouldnt it be nice to speak to the TV to change channels instead of having to search for a remote? (Try explaining to your kids that as recently as 15 years ago, most of us actually had to get up and adjust a knob on the TV.)
- Packaged applications will play a major role in opening up the market place to prospects who werent positioned to pay a quarter million dollar to enhance an existing dual tone multi-frequency (DTMF)-based interactive voice response (IVR) system. Sure there are many pundits who question the viability of packaged speech applications, but while there are real technical and usability challenges for these solutions and its an unproven business model, these obstacles will likely be overcome. The speech market is making its products easy enough to use that even the cynics are beginning to take note of the enhancements in this market.
Challenges remain and we are still not going to see speech software fly off the shelves, as CRM software did in the late 1990s its unlikely that well see anything like that for many years in any market. But with speech vendors finally paying attention to practical issues like cost and ease of use and maintenance, and with the entrance of a couple of serious infrastructure providers that have the financial strength to seriously market these solutions, the speech market is going to take off in the next 18 to 24 months.