Expanding the Role and Capabilities of Search for the Enterprise
by Bill Ives
I recently spoke with Eric Rogge of Exalead about how search is playing an expanded role for enterprises inside and outside the firewall. He mentioned that the major players such as Google, Microsoft, and Yahoo are continuing their battle for broad based Web search. Microsoft has now brought Bing into play in this market.
Exalead is not competing here but focusing instead on the other end of the market, operating in the long tail by enabling businesses to create purpose-built search-based applications. I have written about them before (see Exalead’s CloudView Offers Integrated Search Capabilities). CloudView is one way of attacking the question I raised in my post, We Need to Tear Down the Social Media Silos. Exalead is bringing search to new enterprise tasks and expanding its role through these search-based applications. See my post, Tweepz brings Exalead search to Twitter for an example of this. In our most recent conversation, Eric talked about other search tools they’ve built to find and refine all results from the ocean of Web content. These include audio, video, image-based, and even ranking tools.
For example, they and a systems integration partner built a search-based application for GEFCO, a French logistics company than manages the transportation and delivery of new cars from factory to dealer in Europe. They needed to be able to look at inventory and then keep track of delivery status. GEFCO tried to do this in the traditional manner with a SQL relational database but it was too complex. With a structured database you need to set up multiple data entry forms that require specific data types of data (e.g., car model, location, etc.). There needs to be many data entry forms to do the proper inquiry and car dealers and sales people get frustrated with these specific requirements.
Exalead was able to change the approach for GEFCO to a search based application where you typed in information in a single field as in Google and the search engine did natural language processing of the request without the requirements of a structured database. This greatly simplified the task and gave the car guys what they were used to on the Web. However, in this case you do not want thousands or millions of results as in Google but the single correct answer (e.g., where is the car? When will it arrive?). This application won an Europe award for IT innovation.
Eric said there are many applications, such as this tracking system, with very focused requirements. Another example is patient records in a healthcare system. Exalead also implemented a search-based application for the French postal system to track packages and people. They wanted to be able to answer such questions as where is the package or where is the postal worker? It integrated information from several databases to answer these questions through one interface.
He showed me several other examples. There was the Wikifier (see below), a use case example (wikifier.labs.exalead.com). It parses a web page and creates tags that aid in search. It will also recognize a topic and create a link to a relevant page such Wikipedia article or a search results page from Exalead’s web search tool (www.exalead.com/search). It is a connecting tool or information mashup. As you go through the page you can hover over content and see relevant background information. This same concept could be applied to an enterprise application that covers business topics.

In another example, Eric showed me a restaurant locator tool called ‘RestMiner’. Another use-case example, it mashes together a number of information sources to you get reviews from multiple review sites. A buzz score is calculated from data drawn from multiple sites. Here is another way to break down the information silos on the Web. Google Maps is also integrated to give you location information. Like the Wikifier, RestMiner was built as a proof of concept and could be applied to many business applications.

These examples shown some of the possibilities for search-based applications focused on very specific topics where you want more precise answers. Metrics such as the buzz meter can be build into the application to better aggregate and interpret the data. Many companies have data located in a variety of sources and formats. They do not have to rebuild these applications if a search-based application can work across them breaking down the silos. The Exalead concept of search-based applications has been well received. They had 80 percent growth from 2008 over 2007 and are on track to do it again in 2009. You can also read more about this topic on the Exalead blog.














