Unified information access

Since its launch two and a half years ago, Attivio has been singularly focused on enabling unified access to structured and unstructured information. With Attivio's Active Intelligence Engine (AIE), users retrieve and analyze all types of content and data with simple, search-style queries. New features in AIE Version 2.1 include SQL support (via a JDBC driver), key phrase detection, content spotlighting, entity-level sentiment analysis and integrated connector support for active security. With the release of AIE 2.1, Attivio also provides a platform for building applications, such as dashboards, which provide a comprehensive view across documents, Web sites, e-mails and more, as well as data that has traditionally been limited to database applications.

Last year, Attivio announced a beta program to test its new AIE SQL module that would let customers write and execute SQL queries against AIE directly or from existing enterprise BI tools and applications. The company has since completed the beta with two customers, and AIE 2.1 now includes this full SQL support.

"With the latest version of AIE, we have accomplished some significant milestones. Support for AI-SQL via JDBC delivers the precision of SQL with the fuzziness of search, using a language that is entirely familiar to virtually all developers. At the same time, key phrasing helps to disambiguate an entity based on statistical improbabilities, while our sentiment analysis capabilities can see beyond seemingly positive content to get to the underlying cause of negativity," says Sid Probstein, CTO, Attivio. "These features are not available in legacy enterprise search, data management or business intelligence technologies. AIE brings the best of these solutions into a single UIA platform."

"For information access to work well, it needs strong technical features that support searching, browsing, analysis and decision making," says Sue Feldman, IDC’s VP for search and discovery technologies. "Capabilities like SQL support, key phrase detection and entity-level sentiment analysis are important innovations that provide a deeper level of detection and analysis than have traditionally been available through enterprise search and business intelligence applications. They deepen the understanding of the information to make results more pertinent, and to provide analysis and reporting on top of good information access."

Nick Patience, research director for information management at The 451 Group, says the 2.1 versioning implies that this is evolutionary, not revolutionary. But, he says that of companies offering search-based applications, Attivio is making the best progress and is winning customers at a steady rate. He is impressed with the features included in this version, in particular the support for SQL, key phrase detection and entity-level sentiment analysis.

Probstein says the company constantly works to refine what AIE offers, rather than just roll out new features. For example, AIE has provided entity extraction and sentiment analysis in earlier versions, but now it has entity-level sentiment analysis, which provides a greater depth of insight. Each version of AIE has further enhanced the available security features. The company also plans for additional enhancements to its SQL support, such as sub-queries and case sensitivity. He also showed me several examples of an "active dashboard," which provide a desktop visualization for a specific work environment, such as a sales manager. At this time, Attivio delivers these to customers as specially built professional services applications. But, Probstein says that later this year the company will offer several dashboard modules that can be customized and deployed by customers.

Advance Micro Devices (AMD) recently chose AIE to address a variety of information management initiatives. The initial deployment of AIE is being used to enhance the search capabilities of AMD's public Web site and help manage global Web promotions, including the ability to prioritize query results based on specific campaigns.

"We needed a new approach to information access that extended beyond traditional search capabilities," says Gregg Hansen, VP of applications, AMD. "The relational nature of Attivio's Active Intelligence Engine gives us the ability to formulate queries that understand associations between various content objects to help render results with better relevancy, and also provides the tailored functionality we needed. AIE pulls content from multiple information repositories including our knowledgebase, databases and product catalogs to provide more complete information in response to our users' requests, and helps us better target promotional campaigns."

While the company is pleased at what AIE offered for its Web project (in response to an RFP)--including replacing the two search engines it had been using--it is most excited about the potential AIE offers for connections with AMD's data warehouse, including correlating and reporting. "AIE represents a good jump forward for us," says Hansen.

—Paula J. Hane