Semantic Search, Lexis-Nexis and the Issue of Trust!

It really caught me by surprise when I heard the news that Lexis-Nexis, a leader in legal publishing, chose Pure Discovery, for its patent research search. Again, I want to remind the readers of this blog that semantic search in this context has to do with 'semantics' only - it has nothing to do with semantic web technology as described by W3 standards. Actually, the business need here is nothing more than what you will typically expect an 'enterprise search' use case to be - Basically, getting the right piece of content or document from a huge corpus. In this case, it is approximately 10 million. I am pretty sure that decision makers and decision influencers in this context were well aware of the enterprise search engines like Fast, Autonomy, Endeca and many others in the crowded category of search engines. All three search engines mentioned above have unique algorithms and can handle the scalability aspect easily. So why did they do this effort collaboratively with a relatively unknown company? The main reason cited here is the issue of transparency! Most search engines expect the end user to blindly trust the relevant results - simply trust the algorithm. But in case of PureDiscovery, each query is enhanced by the machine intelligence and "shown to the user" for their complete understanding and engagement. The user can also control the weightage of the query words in a visual manner called querycloud. It seems that search vendors' saying that "its just magic" didn't work this time