Compiled By Tim Stevens Extract relevant technical knowledge out of documents from both internal and external sources with Knowledgist semantic reader software from Invention Machine Corp., Boston. Unlike keyword search engines, Knowledgist discovers concepts and relationships in documents, rather than creating a list of documents organized by subject matter. This desktop tool is first challenged with a problem or question written in common language, such as, "How do you lower cholesterol?" Knowledgist then uses semantic connections in text to create a structured index of relationships, combining a challenge and a solution. These relationships are linked to their original documents, and can be sorted alphabetically or by number of solutions. Knowledgist searches documents in a company database or can be directed outside the company to over 700 Web sites, selected by industry for instance, to extract relationship information and present it as part of the knowledge index.