Speech Vendors: Changing Channels?
Recently I had the good fortune to work with Mark Plakias. And, as everyone probably does with Mark, we got into a spirited discussion (and debate) about the evolution of the speech technology market—and the impact of VoiceXML in the near term. We both agreed that VoiceXML is now becoming a viable alternative as a telephony platform at the enterprise level. We also agreed that the legacy installed base of proprietary, reliable and profitable IVR applications will be difficult to displace. But we believe that enterprises will recognize the benefits of increasing caller satisfaction through easier and faster input using speech, and that will help overcome this barrier, driving adoption.
We were almost seduced into the standards debate by the confusion and uncertainty caused by SALT and X+V. If there are multiple standards then, by definition, there isn’t a standard. Multiple standards did slow our conversation, just as they are probably slowing the adoption of VoiceXML too. Yet enterprises that have to produce tangible results in the near term will likely use the most advanced standard—VoiceXML.
The next turn in the conversation was: "Who benefits from VoiceXML, really?" To date, most of the ASR and TTS engine vendors have sold their software through proprietary IVR vendors. These IVR vendors have the installed base and customers, to which they have been selling newer versions of their proprietary software for years. The ASR vendors used these platforms—and channels—to compliment their value propositions and help drive sales of newer versions of the proprietary IVR platforms that supported ASR and TTS software.
Having these channels in place enabled ASR technology to get into the enterprise. With open standards, including VoiceXML, how will enterprises buy ASR and TTS licenses? There should be a way to obtain these licenses directly from the vendors. Direct sales doesn’t necessarily imply sales people; the licenses could be available through a self-service Web ite, for example.
Mark and I continued down this path, noting that IVR vendors provided several key items on behalf of ASR and TTS vendors. For example, first line technical support and robust license managers are provided by the proprietary IVR vendors. Of course, there is a cost to this infrastructure. And it appears direct sales may increase costs for the ASR and TTS vendors. At the very least it might require a change in technical support and require a stand alone license manager for end-user enterprises.
Proprietary IVR vendors have also gained deep domain expertise from developing voice-user interfaces for the leading speech applications at enterprises such as American Airlines, E*Trade, Fidelity Investments, United Airlines, among others. Their expertise was important in the sales cycle for these complex products, which was a consultative sale for the proprietary IVR platform with ASR functionality. All the ASR vendors provided joint sales teams and pre-sales technical support. Where will the pre-sales function be with VoiceXML? We dug into this for awhile and moved on.
With VoiceXML, will enterprises be able to "un-bundle" their ASR and TTS licenses from the VoiceXML platform? For example, if an enterprise has a VoiceXML platform (open platform), can that enterprise move its ASR and TTS engine licenses from one platform to another? And what if the enterprise is licensing ASR and TTS engines for use with a proprietary IVR platform and then move to VoiceXML? Can the company move licenses across from a proprietary IVR platform to an open VoiceXML platform?
Many alternative possibilities exist, and we played out a few of these scenarios. Where did we end up? A few things seem clear. Specifically, it looks like the channels will be evolving for the ASR and TTS engine vendors. Direct, self-service channels for ASR and TTS licenses may be on the horizon, with license managers of some type and first level technical support. Prices could decrease but the ASR and TTS vendors could get more of those revenue dollars, giving them greater net revenues. (That wouldn’t be without some additional costs.) All this represents a win for the customer, a push for the ASR and TTS vendors, and a loss for the proprietary IVR vendors. It’s also pretty clear that IVR vendors must provide the tools to port existing applications from their legacy proprietary platforms over to VoiceXML—if they want to keep their installed base. If they don't, somebody else will.
Bryan Mekechuk is a partner with Sereno Services and a guest columnist for Mark Plakias in this issue of STM. He can be reached at 408-655-0400 or email@example.com.