Perhaps the speech recognition platform and application business has reached a new level of maturity, as a sudden wave of consolidation has begun to sweep through the industry. The biggest recent move was at the most basic level, i.e., the automatic speech recognition and text-to-speech engine level. While ScanSoft/SpeechWorks has been buying up ASR and TTS vendors for a few years, the acquisition of Nuance was the most significant move by far. Nuance and ScanSoft were far and away the leading vendors in this space. IBM has been in this business for a long time, but they are best known for the use of the ViaVoice engine for dictation applications. There are also a few vendors like Loquendo and Cepstral that have certain areas in which they stand out as the best, but they aren’t nearly as widely deployed as ScanSoft and Nuance. Microsoft has also come a long way in the last few years, but their number of enterprise deployments is still relatively low. Outside of the companies willing to lock themselves into almost completely Microsoft-based solutions, I rarely see Microsoft ASR or TTS being used. Of course, that still leaves them with a lot of department level opportunities and a decent number of enterprise level opportunities.
Even more recently, there were two acquistions that were mostly technology acquistions. First, Microsoft picked up some IP, products, and a few engineers from Unveil. The Unveil website and press releases have never really mentioned any customer deployments, so I’m not clear if they had much of a customer base for Microsoft to acquire. Then this Tuesday, Voxeo bought up the products and customer base from Vocomo, along with a few support personnel. The founders of Vocomo, Danny B. Lange and Mitsuru Oshima, played a major role in the development of the GM OnStar Virtual Advisor product when they worked at General Magic.
The Microsoft purchase of Unveil is a little mysterious to me. As far as I know, Unveil’s products were written in Java and targeted at VoiceXML. Since Microsoft’s SALT now includes most of the functionality available in VoiceXML, it shouldn’t be that hard to covert the Unveil product to generate SALT, but I suspect Microsoft will want to rewrite it in C# anyway. Also, Unveil’s graphical tools were mostly redundant with what Microsoft already has. My guess is that they mainly wanted to acquire the IP. I don’t know anything about the developers at Unveil, so it’s quite possible they had some good engineers that Microsoft also wanted. Generally they try to get all their engineers to relocate to Seattle, though.
Voxeo’s acquisition of Vocomo is much more straightforward. Vocomo was never quite able to make it in the on-premise IVR market. They had a very reasonably priced and quite capable VoiceXML IVR, but it can be tough to make it any hardware market if you don’t have the funding to get big enough to take advantage of economies of scale. Vocomo’s products allow Voxeo to augment their hosted offering with a much more comprehensive on-premise offering.
Okay, so maybe three acquisitions isn’t exactly a “wave” of acquisitions.
Update 11/18/2005: Okay, so Intervoice bought Edify for the surprisingly low price of $33m. Also, it hasn’t been that long since Aspect was picked up by Concerto. Maybe I wasn’t too far off the mark.