Flying Into Intelligent Search
Evaluating Search Tools
As you’re flying toward intelligent search, evaluating the myriad of search tools available is important, just as the pilot’s ability to find SeaTac from the illusory Microsoft office was. Verint Global Consulting Services’ John Chmaj, Sr. Practice Director, Knowledge Management, has four tips for evaluating search tools.
He starts with understanding, which has to do with relevancy. Behind the scenes processing that involves word stemming, synonym association, language patterns, metadata, and contextual analysis operate to respond to a user query with actionable information. Looking for the latest on jets? The search tool should know whether you mean jet airplanes, the New York Jets, or Jet’s Pizza.
Reasoning is Chmaj’s second evaluation point. This is related to understanding, in that search algorithms should have the ability to collate, assess, and prioritize the best responses to the query a user enters into a search box. If it’s jets in the aircraft sense, does the query concern engines, aircraft, regulations, accidents, manufacturing, scheduling, or something else? The evolution of reasoning mechanisms also needs to be considered.
Third is learning. How does the search tool optimize over time based on its analysis of searches performed and results preferred? Machine learning can be applied differently. The depth of leveraging it can vary, and knowing the extent to which it is being applied and how much control and visibility is available are evaluative aspects to pursue.
Interactivity rounds out the four evaluation tips. As user interfaces move from the keyboard to voice and mobile apps, search is now conversational. Tools that interact with users in clear and flexible ways are the new face of intelligent search.
Search technologies are evolving rapidly. Following best practices and evaluation tips will keep you from flying blind. Implementing intelligent search leads to actionable information and insights derived from intelligent search results.