Mobile Speech Recognition as a Disruptive Technology


18 November 2003

I must admit I never did get speech recognition as a disruptive technology. Isn’t it just another dial-pad digit-entry IVR application? Sure, it is much more sophisticated and user-friendly, promising to finally deliver the economic benefits of automation without alienating your customers, and for sure the technology is finally starting to really develop. But is it disruptive in the sense that is disrupts and fundamentally changes the markets for companies products and services? I’m beginning to think that maybe it is.

Adoram Erell from the Intel-DSPC lab in Israel gave a presentation titled Challenges in Speaker-independent Name Dialing for Cellphones at the Speech Technology Seminar 2003 at the University of Hafia, Israel.

The key message on the slide set is this: Bundled with Intel cellular processors.

OK, so it is a small speech recognition application optimized for the single specific task of name identification. But names are no trivial recognition exercise and this is the first time I have seen an embedded voice recognition application without pre-learning embedded in a mobile hand set.

Now it is different. Now we have the beginnings of a technology that is jumping from large, centralized systems requiring specialist hardware and dedicated software to being ubiquitous, built-in to a variety of devices. Today it is your cell phone, tomorrow the TV remote control, the day after that your microwave and your air conditioning system, then your car, office elevator, airline check in … Yes, there are technical challenges, but widespread adoption is usually the stimulus that means we will solve those challenges.

I think I am starting to get it.