I’ve speculated a couple of times in the past that Google’s entry into the “free DA” space had little to do with a business intention to compete with free 411 services, but rather a strategic development initiative that aims to construct a voice vocabulary.
There are many reasons for Google to be in the forefront of a voice-based interface. Nuance’s technology domination (via voracious acquisition) and Microsoft’s capture of TellMe “forced” Google to do what they love to do - build it from scratch. In order to succeed in voice technology, however, you need a huge base of data (voice utterances) to mine into search ontologies. While Google clearly have this in text-entered search, it had no source in voice. Enter GOOG-411. A consumer service on one side, but a massive pipe for capturing large volumes of voice search terms on the other.
As further evidence of this being a means-to-an-end, Silicon Alley Insider reports on the new audio indexing feature announced by Google for Video Search. Cutely/annoyingly named Gaudi, this gives us a more clear sense of Google’s intentions for voice search, and their ambitions in indexing spoken words in video (presumably, podcasts and perhaps music lyrics can’t be too far behind!)
