Nuance Open Voice Search hints at VZW, T-Mobile USA deals: Firm demos voice-based Internet searches

BOSTON – Nuance Communications Inc. is pulling the cloud into the speech-recognition space, and appears to be in the advanced stages of rolling out new services based on the technology with top-tier carriers Verizon Wireless and T-Mobile USA Inc.
Nuance’s activities come as it works to digest a staggering range of acquisitions, valued at more than $1 billion combined, including SNAPin Software for $180 million, VoiceSignal for $300 million and BeVocal for $140 million. And, not surprisingly, the company’s latest business efforts reflect its desire to combine and advance its home-grown technologies with those it has acquired.
Further, Nuance’s moves are aimed at defending against competitors, both small and large. Vlingo Corp. and Yap on the startup end and Microsoft Corp.’s TellMe and Yahoo’s OneSearch on the other are all jostling for ways to make some dough from consumers’ newfound ability to talk directly to their mobile device – and through it to the Internet itself.
“This business has been growing rather dramatically,” said Michael Wehrs, VP of “evangelism and industry affairs” for Nuance, on the sidelines of the Mobile Internet World 2008 trade show here. (Wehrs’ title signifies his near-ubiquitous presence at the wireless industry’s many trade shows, a job he previously enjoyed with Microsoft.)
Wehrs said revenues from Nuance’s speech-recognition efforts doubled this year from last year, to $200 million. That number includes a variety of speech-rec activities, but Wehr said Nuance’s latest effort – Open Voice Search – is an attempt to bring the full force of the Internet to bear in the space.
Basically, Nuance’s new Open Voice Search product allows users to search the Internet by speaking a search phrase. That phrase, spoken into the OVS program installed in a handset, is transmitted through the so-called Internet “cloud” to Nuance’s speech-rec servers, which translate the audio into text and then pass the search phrase to a Internet search engine (Wehrs declined to name the search vendor). The search results are then transmitted back to the handset as a list of Web results.
A cursory test of the service, “What is the name of AC/DC’s newest album?” returned results based on exactly the search term spoken. (The answer, however, was hidden in Wikipedia’s entry for the aging rock band.)
“It’s a hybrid model, is where it’s heading,” explained Wehrs.
Wehrs said Nuance’s speech-rec approach covers both handsets and networks. The company’s speech-rec service can sit inside the handset for address-book lookups (“call Mom,” for example), and doesn’t need to interact with the Internet. This technology is in Nuance’s Vsuite product. However, for spoken Internet searches or e-mail dictations – functions too complex for handsets’ limited processing power – Nuance’s technology must go into the cloud for the necessary translations and computations. These network-based technologies sit in Nuance’s already released Nuance Voice Control product and its pending Open Voice Search offering.
This breadth of products, Wehrs said, sets Nuance apart from its competitors. While some firms focus on spoken interactions with one area or another, Wehrs said Nuance is planning for a day when most or all handset functions can be completed through voice.
However, Wehrs cautioned, “we’re not demonstrating fully integrated products and services yet.”
Though speech-based Internet searches are not new (Yahoo’s oneSearch and Vlingo’s service both offer similar functions), Nuance hopes to cash in on the technology via two mechanisms:
1. Selling the technology directly to wireless carriers, and sharing the resulting revenues with them; and,
2. Integrating the service into already-established, Nuance-powered speech-rec services like address-book lookups and e-mail dictations.
Thus, Nuance is looking to a future where it can make money either through licensing arraignments with handset makers, revenue-sharing agreements with wireless carriers, or combinations of those.
And Wehrs also provided evidence for Nuance’s ambitions. He showed off an unreleased Samsung Electronics Co. Ltd. phone with T-Mobile USA service that could access Madonna ringtones via spoken commands, and a Verizon Wireless phone that could provide driving directions through the carrier’s VZ Navigator service via spoken prompts. Although Wehrs would not confirm Nuance contracts with either carrier, the working demos themselves indicated such announcements may be imminent.