Legal Technology Journal

 
  • Decrease font size
  • Default font size
  • Increase font size
Search - No question – it’s all about answers Print
Written by Oz Benamram, Morrison & Foerster LLP   

A look at how Morrison & Foerster LLP rose to the challenge of making it easier for staff to find actionable information.

Image Just a decade ago, the most significant information problem facing Morrison & Foerster’s lawyers was that of access – simply finding it was difficult. Uncovering data about a specific matter or locating an individual document – even those known to exist within the firm – often involved a variety of information systems, each of which included limited information.

Several years later, the firm’s lawyers are faced with the opposite challenge: they are overloaded with access to too much information. Today they need more than just information. They need answers to questions – relevant information, in context, to aid decision making. This challenge motivated the firm’s recently achieved goal of setting up a usable, unified and context-sensitive search solution that Morrison & Foerster has named AnswerBase.  

Defining the problem

With more than 1,000 lawyers in 19 offices around the world, Morrison & Foerster’s global practice requires its information to reside in numerous systems located throughout the firm:

  • client histories are stored in accounting and CRM systems;
  • documents reside in Hummingbird’s DocsOpen, the firm’s document management system;
  • biographies of members of the firm that describe their specific knowledge and skills are posted on both internal and public web portals;
  • communications, among the firm’s staff and with its clients and others, are contained within the Exchange-based e-mail servers; and
  • files also live on centrally stored and individually accessible file servers.

In the past, users needed to know of, navigate to and query each of these discrete systems, both packaged and proprietary, to find information on specific matters. In addition, with little support for filtering and weighting the results, the traditional document repositories, designed to be good filing cabinets, are not, in themselves, good systems for knowledge exchange. Thus, even when the data had been retrieved, users were further required to sift and manually synthesise the essential results.

For example, an associate who needed to find a member of the firm with expertise on a specialised topic had no direct means to do so other than sending an internal, firmwide e-mail. Instead, they had to serially find documents on matters related to the topic, figure out who had written or been associated with the documents, research the backgrounds of those lawyers, and finally contact the prospective experts.

This situation obviously presented a serious obstacle to the real goal of finding actionable, timely information in context. The first step in formulating a solution was to conduct a study of the needs of firm members and their experience with the existing system. Lawyers and other staff members throughout the firm were interviewed about the kinds and frequency of the questions they sought answers to, the artifacts they sought as answers, and procedures they actually used to search for them. Also a statistical analysis of internal communications was conducted to reveal how members used e-mail, rather than explicit searches, to answer questions (eg the frequency and kind of questions directed to mailing lists versus individuals, the sources and quality of responses to those questions).

The study’s conclusions were unambiguous. As anticipated, users primarily looked for:

  • documents and e-mails;
  • people – internal expertise and external contacts; and
  • information about matters handled for their clients.

But most users found the existing search process to be:

  • time-consuming and distracting – when dealing with multiple systems;
  • unfocused – yielding too many results that could not be meaningfully narrowed;
  • incomplete – as full-text searching did not work across geographically distributed offices; and
  • unreliable – as results did not provide the contextual information they needed to understand and judge their value.

‘Why can’t our internal searches be as simple and fast as Google?’ was a frequent question.

The problems clarified, the question arose as to what search solutions were available to meet the firm’s requirements? The answer, somewhat surprisingly, was not immediately obvious.

Designing the solution

Over the succeeding months, Morrison & Foerster developed a set of detailed requirements for its ideal search solution that addressed all of the issues unearthed in the study, then packaged them into a request for proposals, and sent that out to potential vendor suppliers. In abbreviated form, the criteria were stated as follows:

  • Enterprise search – the solution has to provide one, easily usable point of access to the multiple information sources housed in a variety of platforms throughout the firm.
    Search results should be clustered into logical groups:
    (1) ‘what’ – documents and e-mails;
    (2) ‘who’ – people and companies;
    (3) ‘why’ – clients and matters information; and
    (4) ‘when’ – events.
  • Contextual search – a strong relevancy weighting to rank search results by cross-referencing related information. The search algorithm should take into account the author and the matter that each result is associated with to determine its relevancy. For example, documents that are weighted more heavily in the results should be those drafted by a person who has the search term in his biography; a document drafted by a senior partner; a recent document; or other specific parameters.
  • Expertise locator – the solution must link related information to identify experienced lawyers, inferring the profiles of people from the content of the documents they edited and matters they worked on.
  • Refining results – it must allow users to refine their search results in a logical way, similar to the way they think. For example, in the search results list, clicking on the various source options – ‘e-mail’, ‘portal’, or ‘DocsOpen’ – will narrow results based on their sources. Additional clicking on ‘author’, ‘date range’, ‘department’ etc, will narrow the results even further.
  • Single-click search – the solution must be highly accessible. Search results should be available from any desktop application and from BlackBerry devices. Right-clicking on a word in the message body should initiate a search on that word and display the search results window. Sending a message to a designated e-mail address will initiate a search and return a limited number of results via e-mail.
  • Learning and saving results – the solution should be able to track search activity to identify information needs by user type and to reuse a search to bring up more recent results.
  • Security – the new solution would have to respect private information and retain existing security policies for existing systems. It would also have to support role-based access and other user privileges.

The firm soon discovered from the many responses to the request for proposals that no traditional search vendor had a solution that met its full requirements, even those with highly configurable products.

Consequently, the decision was made to try a different tack: vendors in several relevant software categories were asked successively if they would, in close association with Morrison & Foerster, be interested in developing or extending their existing solutions to meet the firm’s needs.

Traditional legal vendors were approached first, but they were uninterested (or perhaps unable) in devoting attention to the firm’s needs. The firm then turned its attention to online search vendors, but their priorities with their internet-facing business didn’t allow them to support its quest, which they perceived as too specialised. Although social networking vendors did have useful technology for identifying expertise within the firm, they had no solution for searching documents and matters. In addition, these vendors’ systems required users to work in their environments, rather than work in concert with the firm’s existing environments, in order for the solution to work – which was a non-starter.

Asking the right questions

At this point, Morrison & Foerster decided to investigate solutions that effectively managed problems of information overload – those offered by e-commerce vendors. These vendors understood the issue and helped the firm frame its approach in a new way: instead of putting all of the information into a system in the hope that, when someone comes looking for it, they will find it, the team responsible began to think like knowledge brokers – deliver the information that people need right now to make decisions.

Over a dozen vendor responses to the request for proposals were reviewed using a weighted analysis. The firm chose two vendors to test, an established e-commerce search provider and legal search up-and-comer, Recommind. Each developed a prototype of AnswerBase, the firm’s codename for the solution sought.

With two basic working systems in place, 17 focus groups from different offices, practices, and positions throughout the firm were invited to use both systems and report on which one they liked better and why. The technology services group was also asked to evaluate the two systems from a technology and implementation perspective.

In the pilot tests, Recommind’s MindServer Legal ranked higher. Recommind directly addressed the multi-faceted search needs of law firms and developed a comprehensive matters and expertise product to find knowledgeable colleagues and relevant matters quickly and easily. It was obvious that things were heading in the right direction when each of the 17 focus groups was asked: ‘Who in the firm would benefit the most from using AnswerBase?’ Regardless of their role, seniority or location, every one of the groups replied ‘Us!’ With this in mind, a formal purchase agreement with Recommind was signed, making MindServer Legal the engine underlying Morrison & Foerster’s AnswerBase.

Finding the right answers

To address the needs of Morrison & Foerster’s lawyers, AnswerBase was designed with three distinct, interrelated search tabs: ‘Documents’, ‘People’, and ‘Matters’. To provide complete information access to users from across the firm, Recommind indexed the following Morrison & Foerster systems:

  • millions of documents in the firm’s document management system and its knowledge exchange of best practice forms;
  • people information in the firm’s lawyer résumés and HR database; and
  • client and matter information in the firm’s time-notes and its marketing CRM.

By linking time-notes and client information, along with related documents and people information, AnswerBase provides context to search results, and, therefore, a much more comprehensive view for the user.

The system provides three views into its extensive database, allowing the lawyer to quickly access the information most relevant to their query:

  • The system opens by default with the ‘Document’ tab, consistent with the most common use case. However, finding a document is often not enough, or not the most direct path to an answer.
  • By selecting ‘People’ from the top tabs, the lawyer receives a list of the people most knowledgeable about the search terms and a profile showing the information that identifies them as an expert. There is no additional effort required on the part of the lawyers as their work product automatically determines the level of their experience, without the need to manually fill in profiles or update biographies.
  • Selecting the ‘Matter’ tab returns firm matters associated with the search terms, including complete documents, people and time-notes description of the services provided by the firm’s lawyers for that matter.

Each tab has the same features as document search, providing a high level of relevancy and the ability to quickly refine searches using smart filtering.

Using the system

In order to ensure rapid adoption of the system, AnswerBase was designed for intuitive, iterative searches with little training required. The system allows users to search in a way similar to how they think rather than by forcing them to become experts in a new technology product. For example, if a lawyer were assigned to a new matter on the subject of ‘XYZ’, with AnswerBase, the lawyer would take the following steps:

  • Launch AnswerBase in the browser window and type the term XYZ in the search box to initiate a document search across all firm data sources. The initial results could include tens of thousands of documents, since the system has captured everything within the firm’s resources.
  • Results are ranked by relevancy by default, but may be reordered by author, date or other criteria. The relevancy ranking is calculated using a combination of factors: the content of each document; concepts identified within the document; profile of the document’s author; and the profile of the matter associated with it.
  • By clicking on smart filters (see Figure 1, the lawyer instantly receives all written materials based on a specific category or topic. For example, clicking on ‘Client Industry’ and selecting the same industry as his current client will narrow the results to only display documents related to the specific industry. Clicking the ‘Portal’ source button will narrow the results to the ‘Best Practice’ documents posted on the portal. Clicking on ‘Author Group’ will further refine the search to include only those documents authored by partners, associates or staff. Clicking on ‘Author Office Brussels’ will yield only those documents generated in the Brussels office. Each filter reduces the results from tens of thousands to possibly a single-digit figure, as multiple filters can be layered on top of the data set.
Image
  • By selecting one of the search results, the lawyer sees a web page that displays the text of the results, additional context of the results, and available options (see Figure 2). Context is provided next to the document text through a summary that includes detailed information on the document (type, editors), the client (name, industry) and the matter (type of legal services, amount billed, responsible lawyer). Clicking on the author’s name will show that individual’s profile, which will allow the user to contact that person by e-mail or phone. Clicking on the matter associated with the document will display information providing the context in which it was drafted. Clicking on ‘More like this’ will display similar results.
Image
Evaluating results

Morrison & Foerster invited Bruce MacEwen, a legal technology and economics consultant, to evaluate the use of AnswerBase. Tests were conducted using the firm’s summer associates to eliminate the bias of previous exposure to the firm’s older generation of information resources.

The associates were assigned projects that required them to conduct research using the firm’s traditional resources, and subsequently interviewed by the consultant. They then repeated the research using AnswerBase. The three of their experiences detailed in the box above are characteristic of the rest.

To include more experienced lawyers, the scope of analysis was extended to include a wider audience within the firm. AnswerBase worked well for everyone. Lawyers’ research time was consistently slashed with greater confidence in the relevancy and completeness of the results. And across the firm, user satisfaction with their ability to find answers soared. Almost anyone who has used AnswerBase once reports satisfaction and continues to use it.

The bottom line: providing better client service
Image

AnswerBase is now deployed worldwide to more than 2,500 users. The system currently includes all documents, people, and matters within the firm. There are plans to add external contacts, archived e-mail, docket information and portal documents in the near future.

With AnswerBase Morrison & Foerster has dramatically changed the way its staff find answers to questions. More than simply saving time, AnswerBase helps to further enhance the firm’s reputation for intellectual agility and client service. There’s no question about it: enabling the firm’s people to find answers – not just information – helps it better serve its clients. 

Oz Benamram is the knowledge management counsel with Morrison & Foerster LLP, leading the firm’s knowledge management efforts and coordinating its Israel practice.

 

 

Latest News RSS Feed

Advertisement

Advertisement

Advertisement

Advertisement

Advertisement