Smart Search: Business Intelligence and Search

Page 2 of 3

Searching for a Path into BI
Of course, business intelligence vendors aren’t the only ones eyeing the narrowing gap between BI and search. Google has aggressively courted BI vendors to mesh its Google OneBox for Enterprise with their platforms, announcing partnerships with Cognos, Information Builders, and SAS.

Endeca is another search provider elbowing into the BI market with its Endeca for Analytics solution, built on the Endeca Information Access platform. Guided navigation helps users ask relevant questions by putting the results of every search, navigation, or analytic query in a precise context that shows paths to refine and explore further.

And in January 2007, FAST launched its FAST Adaptive Information Warehouse (AIW) to give users direct access to cleansed, fully integrated data with the ability to make ad hoc queries. The company positioned the product in contrast to BI vendors who added a search box on top of a BI tool, saying that AIW “puts the business intelligence solutions on top of the search platform to integrate and orchestrate all of the information needed.”

“FAST AIW offers business users all of the tools needed to unearth new levels of associated information in order to make better decisions,” says Davor Sutija, VP of strategic market development for AIW at FAST. “It also provides the means to detect trends and relationships that were previously never known or explored.”

Accessing Unstructured Data
It seems clear that end users, as well as search and BI vendors, stand to gain from a solution that can combine both structured and unstructured data in query results. But there’s an intermediate step that has to occur before the power of that combined solution is achieved: The unstructured data has to be brought to order so that the BI platforms can process it.

As Business Objects’ acquisition of Inxight attests, text analytics is one such approach. Vendors of text analytics solutions like Clarabridge and Attensity add structure to unstructured content, understanding the hierarchical structure of documents, tagging entities and relationships, and extracting objects such as tables and charts.

Siderean Software takes a different direction to applying structure—it focuses on relational navigation technology. The company’s flagship product, Seamark, uncovers the relationships within and across the data, regardless of format or original source. Siderean founder and CTO Brad Allen says, “Seamark lets end users find and follow relationships between data that may be smeared out across different corporate silos.” He says that Seamark’s ability to present a unified view of all the content in a repository, and to let users drill down into the content in a meaningful way, “allows the web to shine through” the BI platform.

Semantra is another vendor filling the gap between search and BI, though their focus is on relational databases. Company president Davis says, “All our technology is focused on making data easier to find in existing relational databases; we extend what BI platforms do.” Their product, the Semantic Enterprise Platform (SEP), employs semantics, conceptual modeling (or ontologies), and context-sensitive natural language inferencing to provide composite views of already structured enterprise data to enable better decision-making.

Davis says Semantra is careful to not make clients feel like their huge investment in legacy BI solutions have been for naught. “BI vendors like Business Objects have created excellent tools, but they require significant training. We can make it easier to extend the benefit throughout their enterprise,” he says, helping CTOs achieve a better ROI on existing BI resources.

The companies with technologies that sit in the middle of BI and search are hedging their bets by working with vendors on both sides of the equation. Siderean has a reseller relationship with Inxight, while Semantra works with both Business Objects and Cognos. Semantra is also talking to search vendors like Endeca and FAST about collaborating around unstructured search.

Page 2 of 3