As shown in the image above I made a search on google and i had a converstaion with it and gradually narrowed down what I was looking for. Initally it provides with the list of results matching the search keyword by categorizing it into exact match and relative match. It then asks another question like "Are you interested in news/google news, 2005 google, business solutions or engine/engine google?" which helps in narrowing down what we are looking for. In other words, it tries to translate user's thoughts directly into a highly accurate search request. As user clarifies what exactly he is looking for, the results get updated and conversation is shown to the user.
It gives an impression that user is dealing with the engine in natural language and trying to make human computer interaction natural.
Quece is in beta stage and seems quite promising considering the accuracy and the way it deals with the user request. I am not sure how rich is their database indexing, it needs to be tested for many domains and covering many areas. Overall good product and I will defintely watch it as it gradually matures.
Posts
No comments:
Post a Comment