Making Knowledge Management Smarter

By David Southgate

(Back to article)

Enterprise knowledge management (KM) systems were supposed to boost productivity of knowledge workers, saving companies money and, for many companies, KM is doing just that.

But, despite its power, KM has ended up creating its own problems such as making it difficult to find information in vast repositories, and making it bothersome for users to upload new documents.

All of this contributes to the wasted effort KM is supposed to help solve. For CIOs wanting to squeeze their KM's performance, it's important to consider what technologies are on the market and examine what those technologies promise to accomplish.

"One of the challenges today is that companies have different repositories of data all over their environment," said James Segil president and COO of Knowledge Base Solutions, a knowledge management applications vendor. "They might have a database with customer information in it, they might have a Web site that has a whole bunch of content that they publish, they might have an internal intranet -- all of these very disparate data sources need to be searchable and available in a central location.

The research company IDC has claimed that nearly half of the work being produced is partial duplication of something that's already been done, said Toby Redshaw, vice president of IT strategy, architecture, and e-business at Motorola in Chicago.

"The estimates are somewhere between $3,000 and $5,000 per-knowledge worker, (per-year) is wasted effort," said Redshaw, who believes that inefficiency is also due to having too many knowledge workers trying to solve the same problem.

"You can convert that to other untouched opportunities if they just had better tools," said Redshaw.

Thankfully, technologies are emerging to make KM work harder, better, and produce measurable results.

Smarter Search

Take for example Redshaw's experience at Motorola. With almost 4 Terabytes of information online Motorola, Inc. had powerful online knowledge repository, but it had one flaw.

It was easy to set up shared workspace and knowledge repositories on the fly, even across the firewall with partner companies. But when the projects finished and the workers went on to new assignments, the repositories remained hidden in the data soup.

"Unless you knew exactly where the data was or who was on the old team, you couldn't go back and harvest the data," said Redshaw.

To help solve the problem, Motorola turned to Kristian Hammond, a professor in Chicago's Northwestern University Computer Science Department and Director of the Intelligent Information Laboratory. Hammond offered Motorola a way to extend its KM and find those hidden repositories.

Hammond's client-side invention, Watson, lives on local desktops embedded in Microsoft Word, Powerpoint, and Outlook programs. Released in January of 2005 by Intellext, the Chicago company Hammond founded, Watson analyzes an employee's document as its being written, creates an automatic query, goes out onto the KM and pulls back information that can be applicable to the task at hand.

"This has turned a giant sea of information into a goldmine," said Redshaw.

Extending KM

Other technologies are also emerging to make KM more proficient.

At KANA, Inc. of Menlo Park, Calif., the company has developed KANAIQ, a CRM solution that ties into on existing KM repositories and leverages that knowledge with automation that pulls relevant information to customer service agents in real-time.

In a typical call center example, a customer might contact a company with a billing dispute. Normally, an agent would go to a transaction system, bring up the customer's statement, then go to another knowledge base and call up policies around billing disputes.

"If an agent has to go through results sets of hundreds of documents, that's not very efficient," said Brian Kelly, KANA executive vice president of product strategy and marketing.

With KANAIQ the call center agent can trigger the system to search in a number of ways: the agent can type a few words on the screen, the system can be tied into voice recognition, or a decision tree can appear, prompting the agent to ask the customer a series of clarifying questions.

In each of these cases, the CRM system brings back information from the KM system to help the agent to resolve the problem.

Another of the many KM tools now on the market brings in outside data to make workers more aware of a competitive landscape. Factiva, a Dow Jones & Reuters company in South Brunswick, New Jersey, gathers some 9,000 sources -- newspapers, print publications, market data, and newswires -- 60 percent of which are not available on the free Web, according to Karin Borchert, chief product officer of Factiva.

With literally millions of documents in its repository, Factiva knew keyword searches wouldn't be powerful enough.

"We need to understand how do these people work and recognize what's valuable to them," said Borchert. "It's about what problems are they trying to solve."

So Factiva sought to know something about its users -- what part of world they're in, what job they perform, and in what industry. Knowing this, Factiva can bring back articles that will be more relevant to the user and the task they're performing.

For example, the term "Apple" can have wildly different such results for a person working in the computer industry to someone working in agriculture. Factiva takes into account these differences to deliver better search results, said Borchert.

Of course all the technology in the world won't help a CIO if there's no proof of ROI. There are a number of ways to measure ROI in KM. In a customer support setting, CIOs can look at the number of deflected in-bound calls to the call center by providing better self-service tools in the hands of customers. Call reduction time, increased resolution on first call, and decreases in call escalation are also good ROI measures

Training is another metric. If new employees require less training because the information they need is available in a central repository, that is.

"If you believe your company is primarily knowledge workers and that the gathering of data, turning it into information, and executing decisions is the pulse of the company, then this is right at the core of improving the metabolism of your firm," said Redshaw.