Market Update, New Opportunities
Confusion about who the customer is
Who is the customer? Is it the IT manager, call center manager, or telephone systems manager? Actually, it's all of the above. They all have to agree that VoiceXML solutions are a good thing. The problem is that IT departments don't understand how to wire up a telephone patch panel any more than a telephone technician understands how to code a Web application. A coming together of these various branches as well as a unified education process will have to take place. Technical jargon will have to be translated among disciplines and the groups will have to learn to work together more closely than before if voice technologies in the enterprise are going to work. What this tells me is that voice technology in the enterprise will likely have to be accepted and adopted at the CEO/CTO/CIO level before companies will implement it. This means that voice solutions may take more time to sell than say a database system or development tool where the demand and customer are clearer to identify. Over time, selling will become easier as people become more familiar with the technology and the benefits it brings. In the mean time, the market will need to continue the education process.
VoiceXML Rollouts Will Pickup in 2003
There are still many customers who have been working on adding speech recognition to their legacy IVR systems. Right now, I have counted more Fortune 500 companies that are using speech recognition without VoiceXML than with. There will certainly be some large VoiceXML deployments this year within the telecommunications carrier and financial services market, but so far, the majority of deployments and case studies that I have been gathering have been speech recognition without VoiceXML. This is ok--VoiceXML will probably come at the next upgrade cycle for some of these deployments. Some of the companies developing new systems based on VoiceXML may begin rolling them out later this year or early next year. The bottom line is that companies are working on these systems now and we will see deployments begin to pick up in 2003.
SMB Market Underserved = Opportunity
Improvements in ASR and TTS Needed
Despite the great improvements that have been made that enable speech recognition to work on the telephone for production call automation systems, the complexity and time required to develop speech recognition grammars is still too high for the average developer at the average company. In addition, text-to-speech quality is still not good enough to replace recorded prompts.
Here are the two challenges that must be overcome for speech technology to gain the same level of adoption as the Web:
- speech recognition engines must reach the point where they are over 90% accurate without requiring pre-defined grammars
- the difference between speech synthesis output and a human speaker must be indistinguishable.
About Jonathan Eisenzopf
Jonathan is a member of the Ferrum Group, LLC which specializes in Voice Web consulting and training. He has also written articles for other online and print publications including WebReference.com and WDVL.com. Feel free to send an email to firstname.lastname@example.org regarding questions or comments about the VoiceXML Strategy series, or for more information about training and consulting services.
Page 2 of 2