Could you create web widgets that automatically:

• take an webby audio file with spoken word on it,
• run it through voice recognition turning it to text,
• identify key words of your choice in the text,
• tag those key words like in Technorati,
• make those tagged words also HTMLy links,
• identify where in the playtime those tags links exist and insert them into the spoken word file,
• produce a rough table of contents from the indexed words noting where the words of interest first arise and are last mentioned.
• generate a responsiveness within the active spoken word file by making it both keyword searchable and indexable with an introductory table of contents according to the interests of either the maker of the audio file and the listener, and
• maybe even form a rough transcript by identifying the spearate voices and adding a template with their names to form paragraphs?
• THEN...if I was not clear as to my point...revert the data back into the audio file with all this additional interactivity added so, for example, you could link in a text blog post to a specific passage in an audio file?