Broadening the meaning of KM
Ten intelligent information retrieval vendors speak out
Welcome to a mind-expanding tour of available knowledge management technologies as seen from 10 leading vendors that approach KM from different directions.
This article addresses many forms of knowledge, including information residing within and outside of the enterprise, in people and computers. The true value-add of any KM system is measured by advantage obtained for a business process.
A shared focus of all the software vendors interviewed here is attention to the people issues. The vendors aim to serve users, to put information in their hands and to help them do their jobs better by automating knowledge access.
"Right now, corporations are investing heavily in intranets and employees are taking full advantage of them," said Julie Martin, director of product marketing for BackWeb (San Jose).
While every corporation leaps to serve its newly Web-savvy employees, the best strategy is one that doesn't add to the workload.
"People don't have the time to go out and grab the new data, so we proactively provide critical knowledge," she said.
Given that goal, the company has five key tools: polite agent--monitor connection for non-intrusive downloads; remote users--deliver data in packets; network friendly--compress all files; byte-level updating--only refresh changed data where possible; and bandwidth protection--stagger delivery to meet demand cycle.
"We can create knowledge channels, just like we have file channels right now," said Martin. "We have channel profilers that can automatically profile and filter channels. We can set alerts for high-interest info, or apply key word filters for each channel."
Recognizing information overload as a frequent concern with push technology, BackWeb represents an advanced generation of information analysis and filtered delivery. "We communicate politely and we segment and filter so only relevant info goes to each user," said Martin.
"Knowledge retrieval is one of the fastest ROI products," said Mark Demers, director of marketing at Excalibur (Vienna, VA). "Just plug it in and turn it on. You get instant productivity."
According to Demers, "Organizations are lost for a way to jump-start knowledge management."
Referring to Excalibur Technologies' information retrieval, Demers suggested: "Do something about the explicit component of knowledge right now, and install our knowledge retrieval system to tie together access to all of your knowledge repositories."
Vastly expanding the envelope of digitized knowledge and expanding beyond its historic strengths in conventional information retrieval, Excalibur's video analysis engine has been adapted for the Microsoft and Oracle video pumps. "We watch the video for scene changes or certain events, and then we output thumbnails when the changes occur. We make great efforts to eliminate false scene change detections," Demers explained.
The advantage of intelligent software reviewing video is direct and dramatic. If a computer can parse video to the most useful chunks, an important form of information retrieval has been automated. Serial access to such information on videotape is exceedingly tedious. The advantage here is that otherwise inaccessible information can be mined for new knowledge.
According to Ron Weissman, VP of worldwide marketing for Verity (Sunnyvale, CA), corporate power users want two things from search: "They overwhelmingly say, 'Let us build a corporate Yahoo.' And they want to make the same information available differently for different audiences."
Regarding Weissman's first point, the Web-famous Yahoo service is a traditional, human choice index service. Unlike Infoseek or any free text search capability, Yahoo documents are organized in predetermined categories, like a card catalog in a library. The value of such predictable categorization is rapid access to known resources. Compared to "search," such categorical collections are "retrieve."
Weissman's second point is that the same body of information should be parsed into different views and access security based on departments and need-to-know requirements. Like all database functions, access could be determined via log-on and password.
"We need to take the best of document management and apply it unobtrusively. We can't impose a content scheme, so we need categorization systems that learn how users categorize files," Weissman said.
"Speaking as an individual, and not for Verity," Weissman added, "I think XML (extensible markup language) will smooth the way. Instead of indexing and spidering sites, XML will allow much more effective retrieval. It will be the first time that different corporations can cooperate on a common standard, because major software vendors are behind it."
Open Text, www.opentext.com
Abe Kleinfeld, senior VP of marketing at Open Text (Waterloo, Ontario), said, "Information retrieval is an extremely limited view of knowledge management. The process of creating knowledge includes sharing and using it. Context has everything to do with it."
According to Kleinfeld, "The question is how to make KM useful in business life. Expanding beyond information retrieval and document management databases, he suggested another dimension. "Suppose you could learn from discussions that led to decisions," he said. "Instead of searching for words or documents, you could find the tasks that were assigned to solve a problem. You could see the documents that were used, with an audit trail and workflow. You could see how the project schedules actually worked out.
"The best people have the talent to fix the biggest problems," Kleinfeld said, "so you could search all the other projects that relate to this one. You would not only turn up documents, you would find the experts who have confronted and solved these problems before."
From a knowledge management point of view, the system is designed to take tacit knowledge, or to understand the way smart and effective people respond to problems and opportunities, and make that searchable. Kleinfeld said, "Tacit minds leave behind explicit, multidimensional memory that can be absorbed and reused in many ways."
Steve Offsey, director of knowledge management software products at Dataware (Cambridge, MA), started out by sticking a post in the ground: "The U.S. Patent Office uses our stuff to manage 300 GB of text, and we're moving to a terabyte."
"Knowledge is more than documents, the document creation process is part of the knowledge," Offsey said. "Knowledge management is a super-set of document management. Explicit documents are created for specific purposes. There is a lot of knowledge related to the document that is not in the document. The missing knowledge is more important: the audience, why it was written, the business rules at the time, the results. All of this is knowledge that is in people's heads, and in the relationships of documents and data. Employees need to find who knows what. No two people search in the same way. They search differently based on the context of their interests.
"In larger organizations, finding the right people, in time, is the hardest task. If we can figure out who can help, we can share tacit knowledge more effectively. We can create a list of ranked subject matter experts based on the requirements at hand."
"We are applying technology to industrial uses," said Bruce Milne, manager of industry and new media marketing at Fulcrum (Ottawa). "We will contextually summarize documents, so the user doesn't have to read the whole thing." This is a feature that automates not only physical tasks, but the intellectual tasks of reviewing and categorizing documents. It is almost like having the computer read your mail and tell you what is important to look at.
"We have identified targets where our technology can provide a major advantage. One is help desk applications, where we can combine data sources, including E-mail, file systems, Web sites and other sources to deliver cross-repository searching. We can help new people get up and running very quickly by sharing expertise this way.
"Regulatory compliance is another key market," added Milne, "where it is crucial to track multiple data sources, including laws in Canada, the U.S.A., Russia and various states. We can monitor all of these sources of info for relevant knowledge."
Blue Angel, www.bluangel.com
Jon Riewe, VP and chairman of Blue Angel (Valley Forge, PA), explained what Blue Angel software offers knowledge consumers. "All members of the Pennsylvania Technology Council can do an employment search on the Web site. Rather than filling out 42 forms, or logging onto 42 sites, our software lets you search the combined member sites," he said. "That one query will go to all member company sites and provide a consolidated list of responses. Unlike conventional Web repositories, all the information is kept up-to-date by the posting corporations. You are searching live data."
"We automatically generate a Browse Tree," he said. "We offer two methods of discovery: by browse and by search. To automate the Browse Tree, we do all HTML page formatting automatically, presenting up-to-date views of content in dynamically generated pages."
On the Blue Angel gateway demo, users can search 350 libraries to experience the advantage of this meta-level approach to knowledge management.
"Gentia's been great at delivering tools, but now we want to deliver packaged applications," said Michael Gaiss, VP of corporate marketing at Gentia (Wakefield, MA). In an alliance with consultant Renaissance Worldwide (Newton, MA), Gentia is applying proven Business Scorecard methodology to an automated system application.
"Traditionally, the Business Scorecard system has been paper-based," Gaiss explained, "and we have taken the intellectual property of this system and applied automated methods." Stressing the function of cross-level communication, Gaiss explained how the Business Scorecard helps to implement and evaluate corporate strategy on every level. "We identify specific processes that can be improved, and show the correlations of business activities to major goals of the enterprise." he said.
"The Business Scorecard is designed to spot key trends, and focus extra efforts on the most important processes, like customer churn," he added. "We automate the processes that feed the Scorecard, allowing people to employ business measures to current operations, and compare the results to the goals of corporate strategy."
OneSource (Cambridge, MA) offers constantly updated business info feeds, with an intelligent overlay for everyday users. CEO Dan Shimmel said, "We are the only business info suite that offers multiple functions to a user just like an office software suite. We help knowledge workers do their jobs, so we offer tools for prospecting, competitive analysis, trading and so on."
"We deliver a lot of Web links, which we've modeled on the way people work," Shimmel said. "People jump around; it's a non-linear process when people use the Web for work information."
Desktop Data, www.desktopdata.com
"We provide competitive advantage through quick access to relevant information," said Kevin Sperry, sales and marketing manager for Desktop Data (Burlington, MA). "Our traditional users have been very high value info users, people who live in the news. My observation is that this access to information is spreading out through corporations."
"Over time, I see organizations increasingly categorizing external information as a natural resource, as something to be exploited as rapidly and efficiently as possible," Sperry mused.
"We are browser-based, as opposed to Internet-based, since Joe Internet is not going to find our content," said Sperry.
"We are going to be rolling out News Objects, which will make us unique. Developers will be able to build NewsEdge technology into their mainstream applications," he said.
To build a corporate knowledge base, or a shared community of news and intelligence, Sperry suggested, "You can put five top stories up on the corporate intranet home page every day. They can be knowledge about the organization, the industry, the competition or lifestyle of the enterprise."