There was an important illustration of an important business point in the discussion today.
It came from Christina Domecq talking about SpinVox and I was wondering why it was that she would follow this relatively niche model of turning voicemail messages into text messages when, if you’ve got voice recognition technology, you wouldn’t use it for something much bigger, doing away with keyboards and allowing people to talk into their computers.
Her point was that their system is evolving, it’s learning as it goes. It’s listening to all the dialects and voices and different phrases that people are using and the system gets better and better. And her point was that it gets better quicker because it’s exposed to the multiplicity of voices through the telecoms system than it would be in other applications.
The important business conclusion to draw from that is that information is important; it’s worth paying for. And good business strategy is often not just about exploiting what you’ve got, it’s about testing what you’ve got and improving it and gaining information at every point along the line.
What happened next?
There were claims made by BBC News that - far from being innovative technology - Spinvox was merely paying people to transcribe all the words by hand. Spinvox denied this breached the rules of the Data Protection Act, and pointed out that they'd always been upfront that human interaction was part of the mix of approaches to transcription they used.