When you are searching, all you input is the query which is a collection of words. The machine can’t really figure out the context and produces results only based on “relevancy” of search terms.
This often creates a problem when the machine does not understand the actual query. Google launches Knowledge Graph to “intelligently” understand and respond to your query. This is update is being rolled out to US English users and slowly to everyone else.
What this does is, it creates a new block right next to the results summarizing the most popular context. For example, if you searched for Albert Einstein, it will create a box with a picture of Albert Einstein, his birth and death dates, family, works, awards and other related and popular attributes. This will be followed by links to what others have searched.
Below that, there will be other contexts. An example could be, if you searched for Taj Mahal, you were interested in the restaurant nearby and not the monument, it is likely to be found over here.
Its intelligence and human approach combined. Summed up, its a realistic approach to situations and should really work well on saving user’s time and effort.