The matching between social network messages and the web pages is open-ended. We use 2 algorithms in our tool to do the matching, and 1 is trained on wikipedia and recognizes the kinds of things talked about in wikipedia pretty smartly; the other technique
is a context-free matching algorithm on n-grams. The former gives better results and we rank those matches higher. But neither technique is limited to consumer scenarios.
I think the trick to getting this to work for specialized scenarios will be making sure that you have the data you care about available in the social network.