Powerful query technology will optimize Knowledge Management for project managers
published on March 23, 2005
on CNET's TechRepublic.com
Having project documentation available to project managers across the enterprise--whether they are in China or California--would make for better projects and project managers. After all, the more information PMs have, the better decisions they can make in the moment.
One of the promises of Knowledge Management was that it would help project managers centralize and access repositories of company knowledge. In part one of this series, we looked at one project manager who successfully rolled out KM for project managers at his company. Now, we'll examine innovation around KM and search technologies, as well as where KM may go in the future to assist project managers with the continual challenges of bringing projects in on time, on budget, and on schedule.
There's no doubt that KM's ability to centralize archival project proposals, project histories, project plans, and other project artifacts has been a real boon to the project management community.
Better query can improve KM's utility for PMs
But KM hasn't quite lived up to its promise, says Dr. Kristian Hammond, Director of the Intelligent Information Laboratory at Chicago's Northwestern University Computer Science Department.
KM is powerful enough, notes Hammond. It's that KM is sometimes too challenging for users to search for what they need, an obstacle that in general terms can cost companies lots of money. Across the enterprise, for every 1,000 employees who can't find the information they need, U.S. companies spend $2.5 to $3.5 million dollars in lost productivity, according to an IDC report, "The cost of not finding information." To make KM work better for project managers, what's needed is a better query.
But knowing how to build an effective query on the KM can present its own problems. Project managers wanting to create a new project plan, for instance, might not know how to build a query to find a similar document on the KM or even know if such a document exists.
"There really is a gap between a person's ability to articulate what their information problem is and the input that the system receives," says Hammond. Creating the right search string on the company intranet can become like finding the proverbial needle in the haystack.
Hammond has made it his recent life work to develop such a tool. The result was Intellext Inc., a Chicago company Hammond founded, where he helped to invent Watson, a client-side application for the enterprise that automatically creates queries based on the task at hand and brings back contextually related information. The company launched Watson in January 2005.
"What struck us quite some time ago is that information systems, whether search engines, knowledge management systems, or document management systems, the technology behind them is pretty awesome," says Hammond. "And a document that is indexed using fairly low-level statistical techniques can be found if you know how to build the query."
Watson embeds itself in the menu bar of PowerPoint, Word, Outlook, and Internet Explorer and automatically creates a query based on the documents a worker is presently developing in order to go off and find related information in the KM.
Cutting document search down to seconds
In a very early beta version at the Ford Motor Company, Hammond did an integration with Ford's SixSigma Group that demonstrated the tool's utility in helping PMs. Ford has some 8,000 separate information repositories, says Hammond, citing information from a Ford Y2K audit.
"We did an integration with a project management database and we literally were just hooking the thing up and a guy was talking about a project he was on," says Hammond The employee was sure that someone in the company had a process chart that would be helpful for the task at hand. But he couldn't find it anywhere, an understandable situation at a company the size of Ford.
Recounts Hammond: "We asked him, 'Well, where's your project report?' And he pulled up his project report and Watson went off and looked for information and then we said, 'Just type in 'Process Chart' and he did."
The second thing that came up on the list of results was exactly what the Ford employee had been looking for.
"He said, 'You know, if I had had this, it would have saved me two weeks work,'" remembers Hammond.
Future enhancements to KM coming
In the future, Hammond predicts technology will help to build "ephemeral communities of interest," that will continue to save worker time. These communities will appear and disappear to help resolve a particular issue or problem in the moment. To accomplish this, companies will need technology to help them see what kinds of people have touched similar documents, and then locate possible subject matter experts in the enterprise. This would be a major shift from using a slow moving, static database.
"The more we see information systems embedded in the work product, the easier and easier it is to get to the wealth of information and expertise that's related to that work product," says Hammond. "We're about to see a new world where there's far less search and far more information access based on work product."