It is very interesting to see how this new Web2.0 wave allows apparently disparate technologies to be merged together and implemented in radical new ways.

Robert Scoble recently covered how Nexidia is using phonetic search to index video and audio content which then can be searched. And not only that, but by having a textual representation of the content along with some indexing information (such as the position of a certain phrase within a video), they can seamlessly integrate Google AdWords into a viewer and present relevant ads based on the what’s currently being played or listened to.

You can either watch the video interviews provided by Scoble or play with the demo provided b Nexidia.

Now, if there are technologies that allow things such as this to happen, why hasn’t anybody explore its use in the field of self-service? Imagine this. You call into a company’s self service solution and can either be asked an open-ended questions (How may I help you?) or with a directed dialog interaction (Here’s what you can do…). In either instance, and without the need for Statistical Language Models (SLMs) or other advanced recognition techniques, you could be ‘listening’ in real time to what your callers are saying or requesting. Then, based on a pre-defined piece of text, keywords or concepts, you could in real-time drive the interaction by either using those phases as shortcuts, by offering intelligent menu choices based on what they say (not on what you’re offering), and being able to make your system a lot more dynamic by maybe integrating it with a live FAQ back-end or support system that based on things that are taking place in other contact points (fax, email, web, blogs, podcasts, video, etc.) can offer new content/solution over the phone.

Wouldn’t that be sweet?