At its core, knowledge management is about finding lawyers who or documents or matters that can answer another lawyer’s questions about law, transactions, regulatory matters, judges, investors, clients, or industries. With finding so central to KM, software that searches collections of documents and other data sources has long been central to most large US law firm KM initiatives.

To be most useful, software should search across multiple data sources and presents results with the ones most likely to be helpful at the top of the results list. That software is called “enterprise search” or sometimes “federated search.” When search works well, it helps lawyers save time and improve outcomes and it helps firms improve business operations.

For much of this century, one product dominated search in large law firms. Then, a few years ago, many large law firm KM and IT professionals were excited to learn about a product called RAVN, which was enterprise search that used artificial intelligence to enhance its results. Excitement and expectations grew when long-established law firm document management software (DMS) provider iManage acquired RAVN in 2017.

Why the excitement? The DMS is one of the most important data sources in a law firm and, historically, the search built into the DMS was weak. At minimum, the market needs better search for just documents. But more importantly, with the promising enterprise search tool RAVN owned by a leading DMS provider, the hope and promise were not just for better document search, but also search across multiple data sources.

That promise has now been realized at scale, at Sidley, a firm of more than 2,000 lawyers and over 4,000 total users. With Illuminate, Sidley has successfully deployed a large, global, and revolutionary enterprise search platform based on iManage Insight that draws its significant power from connections across documents and other data sources.

CKO Anne Falvey and Director of KM Rich Robbins were kind enough to discuss how what they have internally branded as Illuminate is working, and how they got here.

When did you start planning for this and what was the problem you were trying to solve?

In late 2016, our Knowledge Management Task Force (a group of senior attorneys and staff leaders across various practices and firm departments) endorsed an enterprise search project that ultimately became Illuminate.

We chose that project in support of our Knowledge Management mission, which is to find relevant information and provide access to it, when and where we need it, to improve client service and firm performance.

Sidley is a very large firm with more than 2000 lawyers, more than 2000 staff, 15 global practice groups, 40 practice areas and 20 offices around the world.  Our iManage document management system contains more than 80 million documents and email messages.  Foundation, our experience management platform, includes information on more than 30,000 clients and more than 175,000 matters.

We wanted to make it easier to navigate this sea of information.

When did you complete the roll out and what has reaction been so far?

We conducted a month-long broad-scale pilot in the third quarter of 2019 during which 200 users conducted nearly 7000 Illuminate searches.  After making adjustments based on pilot feedback, we commenced our firm-wide roll out on April 30.  The roll out was completed at the end of May.

The reaction has been positive.   We are seeing broad based use across the firm.  As of today, more than a quarter of the firm has used the system and the utilization trend is very encouraging.  We are getting many requests for group demonstrations and positive feedback.  What we find even more gratifying is the constructive feedback we have received for ways that we can refine Illuminate to make it even more useful.

Are there “war stories” from lawyers and users about how Illuminate has helped them do their work?

Our lawyers and staff have been amazed at their ability to find information that was previously difficult to locate and the ability to find information from a variety of firm data sources.  Finding documents, clients, matters and experience in ONE place, with an attractive intuitive interface is fantastic, but what’s even more impressive is the connections that Illuminate surfaces among data from different platforms.  For example, a junior lawyer looking for a particular type of document to use as a starting place for an assignment can find not only results for precedents but also information about the matter and lawyers connected to those documents, providing context.  One of our favorite quotes from a senior transactional lawyer who had a one on one demonstration of Illuminate was, “Wow.  I am going to live in Illuminate.”  We have had great feedback from secretarial and other staff as well, mostly around how quickly and easily they are able to find information that was buried in various platforms in one place.  In addition, and we could go on and on, we have found that Illuminate is a great place for our lawyers and staff to learn about the firm and each other; summer associates, new hires and laterals use Illuminate to get a sense of who we are and what we do for our clients.

Did the COVID epidemic and work from home introduce many wrinkles?

We swiftly realized that Illuminate would be even more valuable with people working from home.  Therefore, the pandemic served as the final trigger for our global roll out.  We modified our roll out plans ever so slightly to accommodate the fact that we were working from home and then we launched in four waves.

What makes Illuminate different from more traditional enterprise search systems?

A traditional enterprise search system searches multiple systems simultaneously, brings back the results from each and then presents the results in a useful manner, perhaps attempting to emphasize the most relevant items, regardless of where they came from.  Federated search liberates users from the need to learn and work with multiple underlying systems.

Illuminate does that – and more.  Which is where its inherent power comes from.

Illuminate begins by combining the data from the underlying sources into a single search index. It identifies the connections from the information gleaned from the different sources.  Those connections allow users to search the body of information in new ways.

Here is a simple example. 

Assume you are looking for merger agreements for work the firm has done relating to a particular industry.  It is unlikely that the document management system has matter data at that level.  So, while you can find merger agreements for various deals the firm has worked on, quickly identifying those that relate to the industry you are interested in is still going to involve some tedious manual labor. 

What if you happen to have a data set with information about the matters the firm has worked on, and that data set includes the relevant industry for the work?  You can use that data set to quickly identify mergers in the industry you care about, but then you still need to sift through the document management system to identify merger agreements for those particular matters.

With Illuminate, you can do that search in one step.  We take note of the matter level information that we have indexed and use that to supplement the information we get from the document management system, so when we search for merger agreements, we’re able to quickly filter on matter industry to identify those merger agreements that relate to the industry we care about.

As we add more data to Illuminate, regardless of the source, we are able to weave it into the index and that enables increasingly more powerful search.

I know from some of our prior conversations that Sidley was an early adopter of Foundation Software (for experience management) and that you have integrated it into Illuminate.  Can you explain how you use Foundation, why it’s integrated, and the value that provides?

Foundation serves as a vast repository of information about our lawyers, our clients, and the work that we have done for them.  Our work with Foundation came well before we started Illuminate.  While Foundation is invaluable for many use cases, we are going to focus only on the relationship between Foundation and Illuminate here.

One of our goals is to provide a one-stop shop for critical information that people at Sidley require.  We wanted to make it very easy to locate basic information about our lawyers, clients and matters.  In keeping with the spirit of traditional enterprise search systems, we sought to provide a convenient way to access information maintained in Foundation.  For example, Illuminate can easily handle basic client and matter searches because it is powered by Foundation.  We also provide graceful pivot points to allow the user to transition to Foundation if warranted.  At this level, the story is, essentially, federated search.

However, when we cross the line from federated search to a search based on a rich central index of connected data elements, the Foundation database becomes a powerful catalyst.  By drawing the Foundation data into Illuminate and connecting it with data we have from the document management system and other data sources, we make the kinds of connections discussed above.  The beauty of this structure is that because we built a solid pipe from Foundation into Illuminate, as our Foundation dataset grows, Illuminate’s search capabilities grow with it.

Does Illuminate tap sources beyond the DMS and Foundation?

Illuminate also draws relevant data from our firm’s accounting system, people system, and select portions of SharePoint sites that we maintain for various practice areas.  We will add other data sources in the future.

Going back in time, what type of product selection and evaluation process did you follow to select iManage Insight?

We worked with Fireman & Company to run a thorough requirements gathering process during which we interviewed an array of firm personnel including, groups of attorneys from various practices across the firm as well as a group of information technology and knowledge management team members. 

After that, and again, guided by Fireman & Company, we completed a vendor selection process that was based on the requirements established during the prior phase. 

We sought and received proposals from candidate vendors, met with finalists and conferred as a broad team to make our final selection. 

iManage announced the RAVN acquisition at the very end of our selection process.

How has Illuminate compared to your initial hopes and expectations?

Illuminate has met all of our initial hopes and expectations.  It provides the search capabilities we wanted it to have at the outset and as we wanted, it will serve as a solid base for our aspirations going forward.  We are delighted.

Who was on the team that deployed Illuminate and how long did it take?

The deployment team included Sidley, Fireman & Company and iManage representatives.  We could not have deployed Illuminate without the active involvement of Fireman and iManage. 

The core Sidley team included several people from Knowledge Management and Information Technology.  To get a better sense of the team that built Illuminate, and setting aside the core Knowledge Management and Information Technology leadership teams, we had meaningful involvement from all of the following groups within Sidley:

  • Accounting
  • Applications Support
  • Change Management
  • Human Resources
  • Infrastructure
  • Learning and Development
  • Library and Research Services
  • Litigation and Practice Support
  • Marketing
  • Network Administration
  • Procurement
  • Records Management
  • Security
  • Testing and Quality Assurance
  • User Support

We spent the bulk of 2017 completing the requirements, vendor selection and procurement process.  Our design and deployment efforts began in late 2017 and continued through 2018 into the late summer of 2019 culminating in our broad scale pilot.  The system was essentially launch ready late in the third quarter of 2019 but we used the balance of 2019 and the start of 2020 to refine a few things including some items outside of the scope of Illuminate per se but potentially impacted by Illuminate.

Almost all projects of this scale encounter some bumps. What are some you ran into and how did you overcome them?

Enterprise search systems are very large and complex. 

Illuminate stressed our pre-production computing environments in ways we did not anticipate.  For example, at times we found ourselves attempting to design and modify the system in the same environment we were using for testing and quality assurance.  In retrospect, we should have separated those computing environments and we will do so in the future. 

The size and complexity of the system made test and quality assurance difficult; our QA team was heroic.  We will look for more ways to automate their work so that regression testing is easier to accomplish.

A well-functioning search system will expose information readily.  That is what they are built to do.  That also means that if a firm has been relying on “security by obscurity” it will need to make some adjustments.

Do you have a roadmap for future enhancements to Illuminate that you can share?

We anticipate a steady stream of maintenance releases and periodic major upgrades.  The maintenance releases will focus on making adjustments based on user feedback, bug fixes and modest enhancements.  Major releases will be where we deploy major updates to iManage Insight, introduce new data sources and add significant new features.

First and foremost, we will refine Illuminate in response to feedback we receive from the Sidley community, after that, our road map includes, among other things, and in no particular order:

  • additional data sources to further enrich the index;
  • deeper integration with our firm’s intranet and internal practice collaboration sites, including the introduction of an application programming interface (API) that will let us seamlessly embed Illuminate into other applications and platforms;
  • user interface enhancements that will take advantage of capabilities introduced by iManage in releases of iManage Insight that came after the one we used for our initial Illuminate deployment, and providing ways for individual users to tune Illuminate to meet their needs;
  • more comprehensive filtering capabilities driven by the ever expanding universe of data we store in Foundation; and
  • AI driven document classification and data extraction.

I know many firms will be curious about your experience. What are your top lessons that you can share with others thinking about starting this journey?

Do not underestimate the complexity of enterprise search platforms and set expectations appropriately.  Go in with your eyes open.

Our desire to build a sophisticated enterprise search platform was a strategic priority for the firm from the outset.  As such, we were able to get the resources and participation across the organization that was necessary.  We cannot imagine building Illuminate without that level of institutional commitment.

Assess the skills your team will need to design and deploy your system and choose professional services partners who can fill the gaps well.  There is no substitute for experience.

Think carefully about the demands the system will place on your pre-production computing environments, including the volume and richness of data available in those environments.  If your pre-production environments are not suitably rich you will end up discovering things in production that you’d prefer to have discovered and addressed in pre-production.

Testing throughout the deployment process is difficult.  Take the time to develop a thoughtful and repeatable strategy.  Because of the complexity of these systems, changes can have unintended consequences that can be difficult to observe until they manifest in ways that can be difficult to understand.  These difficulties can be diminished with robust regression testing.