Intelligence Built into Intranet Searching

If you could build an ideal information search system, what kind of features would you include?

  • Ability to locate specific information no matter the content format it is stored in?
  • Production of search results that closely match the search criteria?
  • Ranking of information from most to least relevant based on user criteria?
  • Ability to discern content that is not necessarily keyword rich but is relevant to the search?
  • Ability to go deep in an information search and look for content in attachments as well as original documents?

These are important features, but Interact went a step further and added intelligence as a feature. The Interact Intranet software includes an algorithm that actually makes the system responsive and adaptive to user input.

Storing Intelligence

This remarkable feature is found in the software Intelligence Store. When a user performs a search, the system will search all of the content and then produce results ranked on popularity, keywords, relevance, content date, and quality of content, but it can also rank based on user feedback.

In other words, users perform searches and the system captures information as to whether the results were the ones expected. User feedback is assimilated into the Intelligence Store which then refines future searches to produce better results.

Intranet Intranet even has a built-in thesaurus so that keyword searches can include related terms. This keyword analysis tool provides yet more search refinement to insure relevant content is produced.

Filtering System

Yet another feature is the ability to filter the search results based on the type of content the user wants to review. For example, if you only want to read blogs about a certain subject then filter for blog results only.

The intelligent Interact Intranet system improves with each use and that alone is remarkable. But it’s also nice to know that this sophisticated system is also easy to use. It may not be the perfect intranet system, but it has to be pretty darn close.