If there's one thing lawyers have in common, it's that they don't throw a lot away. That's good news when attorneys locate a court transcript or a memo that will save their client's hide in court, but it can make things sticky for those faced with searching through mounds of paper and electronic documents for an errant piece of information.
Although such a scenario seems ready-made for a knowledge-management-based system, law firms aren't generally known for technical daring, and many are still mired in the land of filing cabinets and fragmented islands of information. But as the Washington, D.C.-based firm of Dickstein Shapiro Morin & Oshinsky is learning, expanding the access to institutional knowledge across a law firm makes for a more efficient day in court.
Dickstein Shapiro's efforts to organise its information began in 1992, long before KM became a buzzword du jour. The IT department reviewed the attorneys' document production efforts during a migration to Windows, examining the way the company created and dispensed memos and letters to clients and co-counsel. Based on attorney input, the group decided that the firm needed a way to standardise document creation so that attorneys could concentrate on what went into memos rather than how to format them.
Once that was in place, it made sense to track what documents went to which clients. "We decided it would be a good thing to collect the knowledge that occurs between the user and their clients," says Keith D. Berkland, application development manager for the firm. So in 1996 the firm's IT department introduced a system that records correspondence from an attorney to a client, allowing subsequent attorneys for the same client to review previous work. That meant all attorneys in the firm had access to the same information about a particular client rather than having it inhabit the brain of one or two people. And the seeds of knowledge management began to sprout.
But while the document management system and the relationship management system helped the firm move toward the larger goal of knowledge sharing, it didn't integrate disparate sources of information. Dickstein Shapiro attorneys worked from as many as 30 or 40 different litigation databases as well as paper documents and electronic documents in dozens of formats that stored memos, briefs and transcripts of court proceedings. Users had to learn to use several different applications if they wanted access to all of the information. The system didn't track e-mail, nor could attorneys use it to search through court transcripts and legal briefs. What's more, attorneys fresh from law school had no quick way of learning how to find what they needed. The firm used mentoring efforts to teach newbies, but that consumed both time and money. And the longer it took new attorneys to learn their way around, the longer it was before they could start billing hours on their own.
That's when Angelo Arcadipane, the firm's managing partner, approached the IT department and asked for a system to record and categorize precedence information, which is the knowledge that comes from memos, letters, briefs and opinions related to a particular case. Berkland and Fran Durako, the director of information technology, decided to use that request to institute a company-wide system that centrally stores and integrates information. "To us, [knowledge management] meant sharing and disseminating information in the simplest and most relevant way possible for the users of the information (the attorneys) to do their work," says Durako.
The two decided to use the 125 members of the firm's insurance practice as a pilot group. They chose the insurance group because it works with the most disparate types of information, such as complex insurance statutes that vary from state to state.
Much of the research responsibility for the insurance practice fell to Cherylyn Briggs, insurance research and resources manager for the firm. As the keeper of the insurance group's records and a lawyer herself, attorneys sought her help for case research, which involved examining transcripts, opinions and legal memos from the library's system. Alternatively, attorneys would send an e-mail around asking if anyone knew about a particular case or issue.
Both parts of the system were fallible, however. "[Research] was limited by how much I could read through," Briggs says. For instance, she scanned transcripts looking for particular keywords. If she didn't spot them, the attorney-and the client-were out of luck. "Attorneys could never be confident that they had seen everything we'd produced," says Briggs. And no matter how good Briggs was, attorneys were still relying on the institutional knowledge of one person to cover all their bases. As for the e-mail queries, whether the right person received and read the e-mail was anybody's guess.
So Berkland and Durako gathered two attorneys and five members of the IT department to assess the company's KM needs and find a system that fit. They primarily wanted a system that let users search and index the firm's knowledge stores without having to pull that information from its native environment. So, for example, if a piece of information resided in a Notes database, they wanted a system that could search that database rather than physically move the data to a new environment just so a user could search through it.
In early 1999, after a search that took nearly a year, the review team settled on Fulcrum Knowledge Server from Hummingbird Communications in Toronto. Besides being able to work in native environments, the system allowed for customization. And, says Berkland, the firm already had a good working relationship with the company.
After deciding on Fulcrum, the review team brought in members of the insurance group, including Briggs, and other IT staffers to assess the group's informational needs. Berkland began attending their weekly litigation reviews to get an idea of what documents the attorneys were interested in accessing. Through the information extracted from those meetings and conversations with attorneys, the IT department zeroed in on the two main needs of the insurance group. Insurance attorneys usually looked for either a particular type of document to use as a template, like an outline for preparing a deposition, or information about recurring legal insurance issues, such as what damages are covered under what circumstances or the meaning of certain words in insurance policies.
The group then began to restructure each data repository (databases, Word documents, SQL files) to fit the needs of the insurance attorneys. They also reviewed cabinets full of paper files, which they scanned into the system. The group worked for the first six months of 1999, and the system was initially installed in June. But Berkland cautions that capturing all the data is an ongoing process. "Adding to the knowledge stores is like feeding a hungry elephant," he says. "It never ends."
After doing some testing, Berkland and Durako set the insurance group loose on Fulcrum in September 1999. Users primarily access the system through a web browser with a password, and all of the information is available to each attorney instead of just to Briggs. The system indexes disparate repositories, such as transcripts, opinions, e-mails and legal memos, verifies a user's access and delivers the results in a common screen interface, searching more than 150,000 documents and returning the results in less than a second. For example, a search on a possible witness's name would return any previous depositions the firm had had from that witness; similarly, a search on asbestos and Company B would return anything related.
Look to the Future
There were several inevitable bumps along the way. For one thing, the system didn't record who first authored the documents-crucial to attorneys who want to talk to the author or verify the information. Durako and Berkland added a field to Fulcrum to include that information. E-mail is still not managed systematically: currently, users must manually add e-mail information to a knowledge store that's unique to each client in the system. If they forget, the information is essentially lost.
But while there's no way to make sure every crumb of knowledge gets into the system, Durako and Berkland have witnessed enough success that they plan to extend the system to the other groups in the firm.
For example, Briggs tells the story of an attorney who was preparing to depose a witness from an insurance company. After a search of paper files turned up no mention of the witness, the attorney assumed that the firm had never deposed him. But a full-text search on Fulcrum revealed that the firm had indeed deposed the same witness for a different case. Because that deposition was part of another witness's deposition, it didn't show up in the original search. Without Fulcrum, the original deposition would have remained out of sight.
The IT department plans to expand the system to other sections of Dickstein Shapiro, a slow process that involves an analysis of each group's informational needs. Durako also plans, ultimately, to build an enterprise information portal that will allow cocounsel and clients to communicate with the attorneys. But that's in the future. For now, the goal is simple. Remember the e-mail messages attorneys send around to figure out who in the firm knows something about a particular topic? "We'll have had total success when those messages disappear," says Durako.