Insightful Announces InFact 3.0 For Text Analysis and Relationship Search.

SEATTLE -- New Interactive Text Query Language Enables Much More Specific, Precise, and Accurate Searches than Achievable with Current Search Technologies

Insightful Corporation(R) (Nasdaq:IFUL) a leading provider of software solutions for analysis of numeric and text data, today announced the launch of InFact(R) 3.0, a breakthrough for text analysis and relationship search.

InFact 3.0 delivers a new interactive text query language that enables users to execute custom, flexible, and complex searches from a simple Web interface (much like familiar Internet search tools). Other new features in InFact 3.0 include more precise and expressive linguistic data representations, improved scalability, and the ability to integrate an external "dictionary" of terms -- often called a proprietary ontology -- for rapid customization to address unique problems of specific vertical markets.

InFact is designed to enable professional analysts and researchers, who have long been frustrated by the inaccuracy and limitations of keyword search and early generation natural language processing search, to accurately uncover facts and relationships within text documents. InFact boasts the world's most advanced natural language processing to enable very precise and accurate searching, thereby increasing productivity and the quality of intelligence retrieved from text documents.

Relationship Search and the New Query Language: InFact parses the structure of each sentence, understanding the sentence components, such as which words are nouns and which are verbs. Combinations of nouns, verbs, and other sentence components connote relationships, actions, or facts and are stored in a searchable database structure.

Because InFact is designed to fully understand the grammar, syntax, semantics and linguistic structure of every sentence it reads, InFact's powerful and expressive query syntax can be used to search for very specific relationships, actions or facts. InFact's query language is designed to enable analysts to be very creative in formulating searches that not only search for a particular relationship, action or fact, but also understand action directionality (i.e., company A bought company B, and not vice versa).

In addition, users of InFact can search for relationships involving entity types (e.g., search for any person or organization that was involved in a particular action, not just a particular person), and constrain results based on Boolean combinations and metadata constraints (e.g. the document was published between 1995 and 2005, and the action happened in either North America or Europe). For example, an analyst could search for all cases where one company purchased another company, constraining the search to oil or energy companies, and to purchases that were NOT a hostile takeover, and ones that happened between 1995 and 2005.

"Increased demand to access, analyze, and transform data into actionable intelligence for corporate advantage requires scalable, powerful, and flexible software," said Jesus Mena, author and data mining expert. "InFact offers a powerful combination of relationship search and text analysis that improves the quality of information available for enterprise decision-making by uncovering patterns, trends, and relationships in data that may not be uncovered by traditional search and text analysis products."

"Our customers are reacting enthusiastically to this new release, observing that our product has an unsurpassed understanding of natural language, and that the new query interface is a powerful tool in the hands of their analysts," said Jeff Coombs, CEO of Insightful. "InFact 3.0 is initially aimed at analysts in defense intelligence and bioinformatics, and is currently deployed at a number of intelligence agencies and a pharmaceutical company."

Defense Intelligence

InFact provides intelligence analysts with software to interactively search, quantify, and perform analysis of hundreds of thousands of relationships among individuals, organizations, geographical locations, monetary amounts and all kinds of other transactions within minutes. Using InFact, knowledge workers can perform relationship search and link analysis to uncover cause-and-effect patterns critical for intelligence decision-making.

Pharmaceuticals

InFact's capabilities allow pharmaceutical researchers to gain better insight faster, from volumes of unstructured data, thereby accelerating drug discovery processes. For example, researchers can search and retrieve relationship-based summary reports about all genes, proteins, and pharmacological substances that can block or regulate a specific gene, providing insight that may unlock the next blockbuster drug or stop development of an ineffective drug sooner.

Other applications for InFact include competitive intelligence, financial rating and forecasting, reputation management, patent search and legal discovery, defect analysis, and customer relationship management.

Toppling the Keyword Paradigm: InFact transcends text search as the world has known it over the past two decades. Existing search engines see documents as "bags of keywords" stripped of context or sentence structure. A keyword approach to search and retrieval strips language of context. All information about dependencies, actions and word relationships is lost. The front end is illiterate to user questions. It recognizes the words but can't understand what's being asked. The back end can locate keywords and rank by relevancy but there's no greater comprehension of the facts contained within a document. Traditional text mining and search products typically use a preferred statistical analysis method, such as clustering or Bayesian statistics for linking keywords. The flaw of all these approaches is that context and concepts are destroyed and replaced with a list of keywords ranked by frequency.

By contrast, InFact uses a multilevel approach that uses grammar, semantics, and syntax to capture actions and facts from documents. With InFact, relationships, events and actions are more than mere keyword associations like "word X often occurs with word Y; therefore there is a relationship between X and Y." InFact is designed to perceive events as evidence of some specific activity or attitude, such as "Organization X is hostile to Person Y" or "Gene X inhibits Gene Y."

InFact's reading strategy is based on the most comprehensive language model available. InFact reads every sentence in every document and is designed to develop a deep understanding of each statement it encounters. InFact tabulates the main sense or action of each sentence, determines who is driving the action, identifies the target of the action, and recognizes all entities that are affected or linked to the action. It also identifies modifiers of the action that may add insight, such as the date or location. Using patent pending linguistic data structures, InFact then compares ideas and events across document sources and databases. InFact is able to read thousands of pages in minutes, and to produce a "spreadsheet of ideas" so gigantic and comprehensive that it would take a person a lifetime to assemble by reading the documents.

The Most Advanced Natural Language Processing Technology: Unlike other search tools that have attempted to utilize natural language processing to move beyond keyword search, InFact's indexing is automatic and doesn't require the time and expense of a hand-entered knowledge base (list of terms), manually "meta-tagged" documents, or laboriously prepared rule books. InFact automatically extracts information from every sentence in a large corpus using generalized linguistic interpretation rules. InFact uses intelligent incremental indexing for storing, searching, and manipulating sentence structures. InFact's automated, unsupervised document ingestion, and incremental indexing eliminates labor-intensive maintenance or delays, reading the data that streams into your organization making new information immediately searchable.

Flexible Technology: InFact is designed to ingest millions of documents and update its search database with no downtime. A distributed and flexible architecture can be configured to meet specific organizational needs whether documents reside in legacy database systems, the Web, institutional news feeds, or other enterprise systems. Customers can easily incorporate proprietary ontologies (with simple or multi-parent taxonomies, synonyms, acronyms, and new entity type assignments), endowing the InFact system with knowledge of a vertical domain. A comprehensive Java Search API enables client or server side development, embedding of all InFact functions in other systems, or integration with third-party visualization tools. InFact is built on case insensitive natural language parsing technology, a feature that makes it uniquely robust in defense applications.

Concise Summaries and Interactive Reports Improve Knowledge Sharing: InFact provides interactive relationship-based reports. Users can search for precise concepts and action patterns and then sort results using multiple mechanisms and criteria, such as relationship frequency, geographical locations, or a timeline. Users of InFact can quickly generate cross-document reports and summaries that are dynamically hyperlinked to original source documents, providing an innovative way of navigating document databases. Reports can be shared in a collaborative environment or exported to a central relational database system. In addition, customers can visualize links through a third-party visualization tool.

AVAILABILITY

InFact is available today for Sun Solaris servers. InFact can be deployed on any desktop that can run a Web browser such as IE, Netscape, or Mozilla.

ABOUT INSIGHTFUL

Insightful Corporation (Nasdaq:IFUL) provides enterprises with scalable data analysis solutions that drive better decisions faster by revealing patterns, trends, and relationships. The company is a leading supplier of software and services for statistical data analysis, data mining, and knowledge access enabling clients to gain intelligence from numeric and text data.

Insightful products include S-PLUS(R), Insightful Miner(TM), S-PLUS(R) Server, and InFact(R). Insightful consulting services provide specialized expertise and proven processes for the design, development, and deployment of customized solutions. The company has been delivering industry-leading, high-ROI solutions to thousands of companies in financial services, life sciences, biotechnology, telecommunications, manufacturing, plus government and research institutions.

Headquartered in Seattle, Insightful has offices in New York City, North Carolina, France, Switzerland, and the United Kingdom, with distributors around the world. For more information, visit www.insightful.com, email [e-mail] or call 800-569-0123.

Note to Investors -- Forward Looking Statements

This press release contains forward-looking statements, including statements about the capabilities of InFact (including tasks that users can accomplish with the product); InFact product features; the capabilities of competing products now or in the past; and customer acceptance of our software products. Forward-looking statements are based on the judgment and opinions of management at the time the statements are made. Inaccurate assumptions and known and unknown risks and uncertainties can affect the accuracy of forward-looking statements. Actual results could differ materially from those expressed or implied by the forward-looking statements for a number of reasons, including, without limitation, the risks associated with errors or defects in our software; the risks associated with our ability to compete in the markets we serve; and the risks associated with market demand and customer acceptance of our products. More detailed information regarding these and other factors that could affect actual results is set forth in our filings with the Securities and Exchange Commission, including our most recent Quarterly Report on Form 10-Q. You should not unduly rely on these forward-looking statements, which apply only as of the date of this release. We undertake no obligation to update publicly any forward-looking statements to reflect new information, events, or circumstances after the date of this release or to reflect the occurrence of anticipated events.