The company said the product goes much deeper than similar search products from BI firms like Information Builders, Cognos, and SAS Institute, even though all share the same goal: to make data available to a broader set of business users.
IBM WebSphere Content Discovery for BI combines the best principles of unstructured search with structured data reporting to allow users to access and analyze a broader range of data using a friendly search paradigm. Traditionally these two types of information sources have been managed and analyzed separately, said Marc Andrews, director of strategy and business development for content discovery at IBM, Content Discovery.
Yet nobody wants to use one application to search for BI reports, and a different one to search for other documents, he said. Rather companies are looking for an ability to directly navigate their BI data from their search application and also integrate unstructured information.
Unlike traditional BI products that only analyze and report on structured data typically stored in relational databases, IBM’s Content Discovery for BI can also search across text documents, comment/descriptions/note fields in business applications, Web pages on corporate intranets, and audio and video files.
Content Discovery for BI combines a set of technologies that have been kicking around in IBM’s portfolio of tools around the convergence of structured and unstructured data. Andrews highlighted three facets of the system. First, he said it builds on the capabilities that IBM introduced a while ago to extract knowledge (meaning) from unstructured information in a more structured format so that they can be leveraged by BI systems. This allows users to report on and analyze information buried within unstructured content.
Second is an ability to index BI reports and scorecards in the search engine alongside other unstructured content, which can all be fed into BI systems like dashboards, scorecards, and reports. For example, users can fire a single search that trawls through contracts and invoice databases, and relates that information directly to an analytic report about a particular customer of retailer, which can then be surfaced in a business-user-friendly BI dashboard.
Third, and perhaps the most differentiating element of IBM’s approach, is enabling search against underlying BI data at source, not just data that has been generated in a BI report. Andrews said users could point Content Discovery for BI to crawl through data warehouses and data mart tables directly. We’re providing more ad hoc access and navigation for end users, letting them drill-down and drill-across different dimensions using a friendly search interface for business users, he said.
From a product perspective, Content Discovery for BI packages together several IBM-developed and acquired technologies as a standalone solution. These include: OmniFind search and indexing platform, the open source Unstructured Information Management Architecture framework for connecting different text analytics modules; Information Integrator Content Edition, which is based on content-management technology acquired from Venetica in 2004 for connecting to IBM and non-IBM content repositories; and e-commerce search technology bought from iPhrase Systems last November.
The inclusion of iPhrase’s technology, which was originally designed to help retailers understand and route customer requests and improve their online search engines, is a key differentiator. It allows IBM’s solution to go beyond simple full-text search, and provides a richer contextual awareness for the search query that is based on a semantic understanding of the key concepts, entities, and hierarchical relationships inherent in the data.
For example, Andrews said iPhrase’s technology understands that East could be defined as a sub-region like Mid-Atlantic or that Sales relates to a revenue column in a database. Other search applications simply don’t pick up these hierarchical relationships, he said. iPhrase also maintains industry-specific vocabularies for sectors like telco, financial services, insurance and life sciences that can be tweaked by customers.
IBM’s product leverages UIMA for connecting different text analytics modules that help identify and extract specific business information from unstructured data. The real value of tapping into UIMA is the ability to generate new types of reports by extracting knowledge from unstructured data and data at source and including it in the analysis, said Andrews, who said this capability could also be configured as a service to other business applications.
While other BI vendors have recently added search capabilities to their core BI products, Andrews said WebSphere Content Discovery for BI is the only single solution in the market that integrates the triple threads of enterprise search, unstructured information discovery, and advanced reporting and analytics.
There’s a lot of noise in the market being made around BI and search right now, he said, but we’re doing a lot more than just searching BI content. We’re integrating with broader unstructured enterprise content as well as extracting deeper insights from that content.
Some BI vendors are impressed. Cognos has already integrated IBM’s product with its Cognos 8 BI platform.
