Factiva has developed Fusion by expanding the systems it uses internally to process some 100,000 news articles a day from nearly 8,000 sources. Factiva Fusion's customers can integrate internal content (such as research papers, presentations, proposals, customer records, etc.) with Factiva's news and business information and other external content, including public Web sites, for use in portals and intranets.

"Intelligently managing the volume of unstructured content—85 percent of all data—is a huge challenge facing organizations today," said Clare Hart, president and CEO of Factiva.

The development of Factiva Fusion stems from experience in using component technologies with enterprise customers to determine overall client needs for a more full-service package. "What is new," according to Pat Sabosik, Factiva's vice president of global marketing, "is our new platform of Web services. We have XML across all our product lines with an API and Intelligent Indexing for our customers' content."

Hart said, "Factiva Fusion leverages knowledge assets by providing individuals with relevant information and the confidence that all resources have been consulted." I asked her about integrating licensed content not from Factiva, for example, but from Factiva competitors like LexisNexis or Dialog. She admitted that data from these large services with legacy systems might lie beyond Factiva Fusion's reach. However, she said that many customers preferred access to private publishers and smaller niche market online providers. Hart said they could handle this kind of data source.

Factiva Fusion processing involves the following four components:

• Discovery—Retrieves a company's internal and external (both structured and unstructured) content from multiple sources. Factiva has adapted the AltaVista search engine for Web content harvesting and taps more than 9,000 Web sources in some 20 languages.

• Normalization—Converts all content into a consistent XML format 

• Categorization—Automatically applies a custom taxonomy based on Factiva Intelligent indexing with support from Factiva's partner, Inxight Software

• Export—Presents content for unified searching or downstream XML integration into enterprise applications

The Factiva Fusion system works on Microsoft Windows 2000 Server and Red Hat Linux 7.2 Server. Sabosik indicated that the company already had its technological development team working on Sun Solaris compatibility and would pursue any other platform for which there was customer demand. Some of the components involved in Factiva Fusion stem from partnerships with leading enterprise software application providers, such as Microsoft, IBM, and Plumtree.

Factiva Professional Services, the company's full-service consulting division, works closely with Factiva Fusion customers to develop and implement a content enhancement solution. A Factiva Fusion engagement begins with an Enterprise Knowledge Assessment. Factiva and clients work together to set up a catalog of all knowledge assets, gathering content from file shares, intranets, and premium sources. It will identify critical information needed for specific applications and develop the best content packages to serve those needs. Factiva Fusion also enables document-level authentication and accommodates existing content-permission schemes.

Once identified, the system applies common XML format structure and UTF-8 encoding. Factiva's XML format can accommodate hundreds of data types, which is sufficient to manage the 800-plus formats used by suppliers to Factiva's own content, according to Hart. For example, the system can handle any documents in Microsoft Word, Adobe Acrobat PDF, or Microsoft PowerPoint. Fusion converts into Unicode, normalizes content in any language, and improves searchability by removing the ambiguity caused by accents and punctuation.

To process the content, Factiva Fusion develops a custom taxonomy using Factiva Intelligent Indexing as the underpinning. Built originally by experts from Dow Jones and Reuters, it provides a basis of 300,000 daily-updated company codes, as well as a schema for common industries, news subjects, and geographic regions. Updates to the overall Factiva taxonomy are quarterly. Companies that do not want to build on Factiva's established taxonomy can use their own. In developing and testing the categorization rules and document collection automation, the Fusion teams develop "training sets" to assist the automatic categorization software in assigning accurate taxonomic terms to content. The final step is the integration of content into work-flow applications.

Charges for Factiva Fusion are negotiated based on the amount of work, resources, and talent required, according to Hart. For more information, go to http://www.factiva.com/fusion.

Hart stated that the company is marketing Factiva Fusion to corporations and government agencies. "We usually sell to three kinds of people: content (e.g., information professionals), technology, and business buyers. Implementation arrangements rely on senior technology personnel. They handle security dealings, for example."

Hart said that at the taxonomy-development stage the company preferred, if possible, to work with in-house information professionals, though it does offer full-service contracts that cover all personnel required from the Factiva Professional Services group. At this point, Hart said, Factiva has no program to train information professionals in Factiva Fusion implementation, but she did mention Factiva's long-standing close relationship with SLA and suggested that it might want to develop a program with that organization.

The biggest change for Factiva, according to Hart, was the move to a service role. "It's a new world. We're not just in content anymore. We're selling services and technology."