Find the expert

Many business and technical problems are best solved not by retrieving electronic information but by retrieving a human expert. Finding expertise within an enterprise can be one of the most vexing problems in knowledge management. Often, managers or staff members suspect that the expertise is available, but they don’t know where to look.

Early this year, Tacit announced its “expertise automation” technology and its ESP (expertise search for portals) software. ESP allows companies to put a component within their portal that enables searching for people and expertise. Any documents, files, etc., that have been published in the portal can be analyzed to build expertise profiles of all people whose work is represented withn that environment.

The Lotus Knowledge Discovery System, which includes the collaborative knowledge portal, K-station and the Lotus Discovery Server, automates the process of identifying expertise by monitoring the activity of documents, e-mail and other files.

“We assumed that experts are too busy to create and maintain profiles,” says Scott Eliot, director of KM product strategy at Lotus Development. “We believed that the information that helps identify experts should be derived from material already available.”

The Knowledge Discovery System analyzes patterns of authorship and use. The K-station portal serves the usual function of aggregating content, but also brings in Lotus’ traditional collaboration tools such as Sametime messaging and the QuickPlace virtual meeting space.The ability to locate experts is an example of what the Knowledge Discovery System can do and represents a way of tapping into corporate expertise. The system can also provide added value to knowledgebases by clustering documents from multiple repositories into a single category and creating topical knowledge maps.

Yet another set of relationships can be derived, identifying people whose profiles and document activity indicate a relationship to a particular topic. Lotus refers to this process as “affinity mapping.” That information can be shared throughout the enterprise, allowing development of additional collaborative relationships.

Verity Verity (verity.com), well-known for its widely used search engine, has also taken on the challenge of accessing human expertise by identifying “social networks.” The newly introduced Verity K-2 Enterprise is a portal infrastructure that automates that process. Document and e-mail activity is monitored for the purpose of inferring relationships and expertise.

“The value of the data is not in bits and bytes, but what it means to people, the company and the business,” says Ashok Chandra, senior VP of development and new business activities.

Verity sees the next generation of portals as consisting of three levels: (1) search, applied when the user is trying to locate a specific document; (2) organize, to find relevant information in a body of knowledge; and (3) social networks, in which people interact with relevant information and communities. The Verity K-2 Enterprise portal is designed to support all three functions.

As might be expected from Verity, the search capability is strong and sophisticated. For example, a parametric search allows access to structured and semistructured information by using metadata and setting criteria for information retrieval. The Verity K2 spider indexes documents automatically across all repositories. Documents can be organized by an intelligent classification system based on business rules combined with a state-of-the-art automatic classifier.

The system gradually learns to identify documents as examples of a certain category. Adaptive results ranking allows the relevancy to be modified based on user feedback. In its decision to incorporate social networks into its analyses, Verity is recognizing that information is not intrinsically relevant but attains value from its use by communities of practice.

In a move that reflects both old and new concepts, QUIQ (quiq.com) has developed a software product that allows “mass collaboration,” in order to tap into the expertise held by customers themselves. Reminiscent of CompuServe forums and computer group bulletin boards in which users traded information, QUIQ adds an organizing infrastructure that creates a searchable knowledgebase. Compaq Computer (compaq.com) announced in May that it was implementing QUIQ to help its customers share knowledge through its Compaq Customer Communities.

“The information managed by QUIQ is organized in a customer-centric taxonomy,” says Kartik Ramakrishnan, co-founder and VP of business services at QUIQ, “but the system also carries metadata about the information, the people creating it and their roles.”

Data mining tools can be used to analyze the metadata, and the results can then be incorporated into the search framework so that the search criteria become dynamic. The resulting search accesses both text and structured information to guide the user to the appropriate expertise.

Judith Lamont is a research analyst with Zentek Corp., e-mail jlamont@sprintmail.com

KMWorld Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues