by Edward Prewitt

Assessing Knowledge Management Products

Feb 15, 20018 mins
Enterprise Applications

Knowledge management is the Holy Grail of the modern company, much rumored but rarely found. Many software solutions?from document management and data mining to search engines and portals?have claimed to be the key, only to fail. “Knowledge management is a mix of all those disciplines with a good dose of know-how on top,” opines Elise Olding, vice president for knowledge, e-learning and collaboration at the Hurwitz Group, a Framingham, Mass.-based consultancy. In other words, knowledge management can’t be found in a shrink-wrapped box.

And yet some recently released software products seem to make promising progress toward nearly transparent knowledge management. While these apps take very different slices at the problem, they share an elegant approach to structuring information. Rather than asking users to create time-consuming data taxonomies (the process of coding, tagging and distilling keywords from the reams of information inside an enterprise), Knowmadic’s KM.Studio, Autonomy’s peer-to-peer capability and KVS Software’s Enterprise Vault all attempt to organize information effortlessly and invisibly. They operate in the background, silently expanding the corporate knowledge universe?even incorporating information from beyond the company walls.

In short, these products claim that they can help your users become a collection of knowledge workers without changing the way they do anything?and that simplicity is the key. “Knowledge management needs to be painless,” Olding says. “Until we get to the point where technology mirrors the way people work, knowledge management isn’t going to happen.” KM.Studio, Auton-omy’s P2P technology and Enterprise Vault allow workers to structure and restructure information?from internal databases to webpages to corporate memos?on an as-needed basis instead of forcing all employees to become archivists.

Canvassing the Web

“You cannot get all information from one source?but Knowmadic can,” says Dante DeWitt, CIO of Equinix, a Mountain View, Calif.-based company that manages secure operations centers for e-businesses. DeWitt says that not only is the product ideal for organizing thousands of data sources, it works the way users need it to. “Each individual can use the tool and configure it his own way. It’s like having your own personal librarian.” Santa Clara, Calif.-based Knowmadic’s KM.Studio, which started shipping last October, combines a sophisticated Web browsing capability with automated work processes. The product manipulates both the Web and enterprise applications like a database, says Knowmadic Vice President of Business Development Brian O’Neill, thereby letting users find, extract and manipulate information in a variety of formats (such as HTML, XML, PDF or Open Database Connectivity-compliant databases) from nearly any source.

A simple drag-and-drop interface lets users connect navigation and extraction objects to business rules they create. The users can then record their actions and set KM.Studio to repeat the process. This macrolike capability is very useful, says DeWitt. “You can decide, for instance, that you want to see all analyst reports related to a certain sector every day and do it simply and graphically.” DeWitt used KM.Studio while CIO at Chase H&Q, the investment banking arm of Chase Manhattan Bank, and he is now considering buying the app for use at Equinix.

Installing and running KM.Studio need not be a chore. Although nontechnical users generally turn to the IS department for set up, O’Neill says, sophisticated users can configure KM.Studio themselves in about a day.

Among Peers

Cambridge, England, and San Francisco-based Autonomy has gained repute since its launch in 1996 as the only company with software based on the work of Thomas Bayes, an 18th-century English cleric who developed probability estimation techniques while searching for mathematical proof of the existence of God. Autonomy’s technology uses probability theory and pattern-matching algorithms to identify similarities in unstructured information?text, spreadsheets, PowerPoint slides and other material that people can understand but computers usually can’t process. In November, Autonomy announced a soon-to-be-released peer-to-peer version of its technology that extends the capabilities of its basic platform, linking employees’ PCs so that Autonomy software can automatically search them for parallels. For large, far-flung companies, this new function can harness existing knowledge?letting em-ployees put it to use instead of having it lie hidden and unused?the fundamental definition of knowledge management.

“When I set up an agent on Web servers, it immediately came up with another person in the company who’s interested in Web servers,” says Krister Lagerborg, manager of the Business Intelligence Center unit of Ericsson in Stockholm, Sweden. “Autonomy links to other people with similar agents and generates a list.” That list then becomes a valuable source of information that might otherwise have gone untapped. Lagerborg’s group launched Autonomy’s Portal-in-a-Box product in February 2000 for use with searches of the business intelligence portion of Ericsson’s intranet. The news and analysis content of the business intelligence portal make it the second-most visited site on the company’s intranet, with about 10,000 regular users (out of 105,000 Ericsson employees globally).

Because Autonomy’s search capability is extremely fast?hundreds of gigabytes of content within tenths of milliseconds?the platform can tackle each query afresh, searching for similarities among huge amounts of information without the need for tagging data. Lagerborg has found Portal-in-a-Box faster and more reliable than the traditional search engines it supplanted. Autonomy’s added features, such as automatic e-mail notification and search agents, are another selling point. About 3,000 users at Ericsson have created agents to automate searching the business intelligence portal, Lagerborg says. Autonomy generates the peer-to-peer linkages automatically.

This background operation?in personalization, profiling and notification?is a principle of Autonomy, says David Appelbaum, the company’s general manager for North America. “Any kind of system that forces people to do things differently than they already do is going to be doomed,” he says. This ease of use extends to installation; Autonomy’s platform is highly generic and modular. IS staff can configure the platform any number of ways. Ericsson devotes only one IS worker to running Portal-in-a-Box.

It’s in the Mail

Knowmadic and Autonomy reach out to the Web and other external sources for their information. In contrast, KVault Software (KVS) focuses on an often-overlooked internal repository of knowledge: e-mail. “If you use e-mail, you’re a knowledge worker,” says Nigel Dutt, CTO and founder of KVS, headquartered in Winnersh, England. Yet unlike corporate databases and other group information stores, e-mail is rarely managed. Instead, senders and recipients, reluctant to share e-mail, retain the files on their hard drives.

Researchers retained by KVS estimate that corporate users across the globe send 4 billion e-mail messages a day. As most anyone who uses e-mail can attest, much of that influx piles up, waiting to be read and replied to or deleted, all the while scrolling farther down the inbox. Dutt founded KVS in December 1999 after realizing that e-mail had become a de facto long-term storage repository. Yet limits on server space typically force IS departments to prod users to clean out their mailboxes every few months. When that happens, Dutt says, “people either read all those e-mails and do something with them or?what usually happens?they throw them out. That information is lost. And when people leave the company, knowledge is lost, because there is no record.”

KVS’s solution is Enterprise Vault, a generic, scalable application that typically resides on a Windows NT server. Enterprise Vault automatically archives all e-mails in a company’s Microsoft Exchange or Outlook system, allowing IS staff to slim down server space while giving users instant access to all e-mails. “Users put an icon on their desktop and retrieve archived e-mail themselves. It’s that simple,” says Daniel Rodriguez, former vice president of corporate e-mail at Donaldson, Lufkin and Jenrette in New York City. Employees needing to retrieve old e-mails no longer have to go through his office, Rodriguez says. “It’s a big time savings for users and for [IS]. They didn’t have to call the help desk, and for IS, we don’t have to go to tape.” Before installing Enterprise Vault, Rodriguez’s office would often have to download tapes to find old e-mails?a process taking seven hours per tape.

Of course, better archiving alone doesn’t advance knowledge management; the problem of unmanaged information remains. Where Enterprise Vault improves on the standard e-mail package, however, is in its search capabilities. Enterprise Vault incorporates AltaVista’s search technology, which users can configure to search the headers and bodies of all stored e-mails, either for individual users or systemwide. This enhanced access to e-mail is a potential boon to companies that are poor at sharing knowledge. “There’s an incredible amount within e-mail that’s critical to the way companies now do business ,” says the Hurwitz Group’s Olding.

For all the attention paid to knowledge management, Olding dislikes the term, finding it ill defined. “If we think about the times we’ve learned and had ’aha’ moments, that’s not managed,” she says. “Simply put, I look at knowledge embedded in the process.” Technologies like those profiled here “are going to be the ones that survive,” she says, “because they work the way people do.”