According to market researcher Frost & Sullivan, the UM pot will grow from last year's $549 million to $5 billion by 2005. The Polaris Group is even more optimistic, estimating a $6.3 billion take by 2004. But what's a billion here or there among friends -- the point is, everywhere we look, UM is once again being touted as a market with huge growth potential. We're not saying we're surprised by this. After all, any technology that promises to manage the deluge of electronic communications we face daily is a natural to attract hype. We are, however, a bit skeptical about this newest round of lofty predictions, because past prognostications in this space have proved to be wildly optimistic. So what's different this time around?
In a nutshell: working smarter, not harder.
First, the technology and infrastructure continue to mature, enabling UM products as well as complementary technologies like VoIP (voice over IP) to consolidate around standards and develop synergistic relationships. Digital convergence is the name of the game these days, and UM is riding the coattails of a major push to bring multimedia to the Internet and voice to the data network.
|
What Do Readers Think?
Check out our e-poll results on unified messaging.
|
Second, mobile workers want to be able to access, respond to and manage all forms of communication from all types of devices, including landline phones; cell phones, both with and without Internet capabilities; PDAs; and notebook computers. UM is a good example of a technology that is moving from nice-to-have to must-have status. To remain competitive, we need to manage that daily communications avalanche efficiently. This level of efficiency is possible, however, only if all our vital information is in one pot, with that vessel reachable by any device.
Because we're talking about access to any type of communication from any device, TTS (text to speech) and speech recognition are the two key technologies to consider. Soon we'll be listening to and replying to e-mail over the phone, eventually managing all our messages via voice command.
At first glance, TTS technology would seem to be the easier of the two technologies to implement. After all, speech-recognition products face a mountain of nuances inherent in the spoken language, including different dialects, slang and regional accents. Sometimes speech recognition during face-to-face conversations between humans speaking the same language hits only the 50 percent mark; no wonder it's a difficult proposition for a machine.
However, first looks can be deceiving. TTS has proved to be almost as tough a nut to crack as speech recognition because a voice that doesn't sound human just does not cut it. Also, when you consider the informal nature of most e-mail -- all those abbreviations, diagrams and emoticons built out of punctuation, URLs and tables, and missing or misspelled words -- it's easy to see why TTS vendors have their work cut out for them. Want to hear for yourself? Visit demo sites by Lernout & Hauspie Speech Products or SpeechWorks International, and enter some typical e-mail text.
Although you'll find that today's TTS offerings sound more like Colossus from The Forbin Project than HAL from 2001: A Space Odyssey, products like SpeechWorks' Speechify and Lucent Technologies' Articulator Text-to-Speech Software Development Kit are helping to advance the state of the art, enabling resellers, third-party application developers and ASPs (application service providers) to introduce applications that are usable, rather than just niche items.
Even in the midst of this progress, however, we did get a good laugh from the TTS product that recently read to us, in a synthesized voice, an e-mail subject line guaranteeing us a larger penis, only to be followed by the next subject, offering a free pair of panty hose. We're so confused -- but that's another topic for another time.