Is This the Next Generation of Search?
-- Library Journal, 11/7/2006
"We view the need to go to a search engine to find information as a failure," said Jay Budzik, chief technology officer of Intellext, which produces the new Watson search application. "Your computer should anticipate your needs. If we succeed, this will be one of the standard ways to receive information." Watson analyzes the context of what users are doing on their computers and proactively sends them relevant information from search engines, research services, corporate intranets, and other content sources. It was introduced in 2005 for corporate clients. "They had content in their portals that no one was using, and a variety of subscription resources that no one was using," Budzik said. Last summer a free ad-supported version emerged.
The company has just started to talk with library vendors, both metasearch companies and ILS vendors, since "there is a clear opportunity for libraries to leverage their investments in premium." Wasn't metasearch invented to help libraries leverage expensive proprietary content? Yes, said Budzik, "but the problem is that, with a separate destination, you're competing to change the user's behavior and shift their attention."
Some searches are driven by user events, such as opening a document or move to a new URL. Budzik said Watson "also monitors whether you've significantly changed the document you're working on." Then "precise queries" are sent to search engines. And what if the user thinks results are lousy? "We make it easier to pick part of your document and say, "Hey, I really want to focus on this. Refine it and the results are related. We think this is a key point in moving search forward."



















