Speech recognition going mainstream – or not?

Interested that the Gartner hype cycle for Emerging Technologies, shows Speech Recognition is coming out of the slope of enlightenment and moving towards mainstream adoption, albeit in 5-10 years. Speech recognition seems to be a technology that is perpetually less than five years away from being adopted as mainstream and still has a long way to go.

Speech Recognition technologies are hard enough to work in call centre applications where callers are asked to provide very specific information – my experience calling credit card company lines is that they seem to have enormous difficulties even recognising the numbers 0-9 although this might just be that I have an inpenetrable accent or something. This seems to be where they can have the most immediate impact however.

Using technology to recognise and interpret more complex information however is another matter entirely as the recent Spinvox debacle demonstrates.



Possibly the technology industries worst kept secret – Spinvox uses people to transcribe lots of the messages that it transcribes from voice-to-text. No great surprise although the extent to which their investors have been led to believe that this was a technology vs people-powered solutions seems to be a case for speculation. Still, any investor putting money (and £120 million of cash registers on any VC radar), would surely have done some proper due diligence wouldn’t they…?

Reblog this post [with Zemanta]